From a6a444d7e9e37b99faa28dc5a5611a2055c841aa Mon Sep 17 00:00:00 2001 From: Simon Fan Date: Mon, 9 Dec 2024 12:40:00 -0800 Subject: [PATCH] Mark everything in CompileId as optional, add compiled_autograd_id --- src/lib.rs | 19 +- src/parsers.rs | 19 +- src/types.rs | 23 +- tests/inputs/cache_hit_miss.log | 6033 ++++++++++++++++++++---------- tests/inputs/chromium_events.log | 453 +-- tests/inputs/comp_failure.log | 187 +- tests/inputs/simple.log | 2212 ++++++++++- tests/integration_test.rs | 24 +- 8 files changed, 6470 insertions(+), 2500 deletions(-) diff --git a/src/lib.rs b/src/lib.rs index 94017d1..c082889 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -381,10 +381,22 @@ pub fn parse_path(path: &PathBuf, config: ParseConfig) -> anyhow::Result = @@ -450,7 +462,10 @@ pub fn parse_path(path: &PathBuf, config: ParseConfig) -> anyhow::Result { .map_or("(unknown) ".to_string(), |c| format!("{cid} ", cid = c)); let mut cid = compile_id.clone(); if let Some(c) = cid.as_mut() { - c.attempt = 0; + if let Some(_frame_id) = c.frame_compile_id { + // data migration for old logs that don't have attempt + c.attempt = Some(0); + } } let stack_html = self .stack_index diff --git a/src/types.rs b/src/types.rs index 8baa23b..361b403 100644 --- a/src/types.rs +++ b/src/types.rs @@ -128,16 +128,27 @@ impl StackTrieNode { #[derive(Eq, PartialEq, Hash, Deserialize, Serialize, Debug, Clone)] pub struct CompileId { - pub frame_id: u32, - pub frame_compile_id: u32, - pub attempt: u32, + pub compiled_autograd_id: Option, + pub frame_id: Option, + pub frame_compile_id: Option, + pub attempt: Option, } impl fmt::Display for CompileId { fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - write!(f, "[{}/{}", self.frame_id, self.frame_compile_id)?; - if self.attempt != 0 { - write!(f, "_{}", self.attempt)?; + write!(f, "[")?; + if let Some(compiled_autograd_id) = self.compiled_autograd_id { + write!(f, "{}/", compiled_autograd_id)?; + } + let frame_id = self.frame_id.map_or("-".to_string(), |v| v.to_string()); + let frame_compile_id = self + .frame_compile_id + .map_or("-".to_string(), |v| v.to_string()); + write!(f, "{}/{}", frame_id, frame_compile_id)?; + if let Some(attempt) = self.attempt { + if attempt != 0 { + write!(f, "_{}", attempt)?; + } } write!(f, "]") } diff --git a/tests/inputs/cache_hit_miss.log b/tests/inputs/cache_hit_miss.log index 9fe0f4f..c931278 100644 --- a/tests/inputs/cache_hit_miss.log +++ b/tests/inputs/cache_hit_miss.log @@ -1,69 +1,74 @@ -V1003 10:10:50.629000 2235078 torch/_logging/structured.py:22] {"str": ["/data/users/oulgen/pytorch/torch/_dynamo/convert_frame.py", 0]} -V1003 10:10:50.630000 2235078 torch/_logging/structured.py:22] {"str": ["/data/users/oulgen/pytorch/test/inductor/test_codecache.py", 1]} -V1003 10:10:50.630000 2235078 torch/_logging/structured.py:22] {"str": ["/data/users/oulgen/pytorch/torch/_inductor/test_case.py", 2]} -V1003 10:10:50.630000 2235078 torch/_logging/structured.py:22] {"str": ["/data/users/oulgen/pytorch/torch/_dynamo/test_case.py", 3]} -V1003 10:10:50.631000 2235078 torch/_logging/structured.py:22] {"str": ["/data/users/oulgen/pytorch/torch/testing/_internal/common_utils.py", 4]} -V1003 10:10:50.631000 2235078 torch/_logging/structured.py:22] {"str": ["/home/oulgen/.conda/envs/py311/lib/python3.11/unittest/main.py", 5]} -V1003 10:10:50.631000 2235078 torch/_logging/structured.py:22] {"str": ["/home/oulgen/.conda/envs/py311/lib/python3.11/unittest/runner.py", 6]} -V1003 10:10:50.631000 2235078 torch/_logging/structured.py:22] {"str": ["/home/oulgen/.conda/envs/py311/lib/python3.11/unittest/suite.py", 7]} -V1003 10:10:50.631000 2235078 torch/_logging/structured.py:22] {"str": ["/home/oulgen/.conda/envs/py311/lib/python3.11/unittest/case.py", 8]} -V1003 10:10:50.632000 2235078 torch/_logging/structured.py:22] {"str": ["/home/oulgen/.conda/envs/py311/lib/python3.11/contextlib.py", 9]} -V1003 10:10:50.632000 2235078 torch/_logging/structured.py:22] {"str": ["/data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py", 10]} -V1003 10:10:50.632000 2235078 torch/_dynamo/convert_frame.py:915] {"dynamo_start": {"stack": [{"line": 916, "name": "", "filename": 1}, {"line": 14, "name": "run_tests", "filename": 2}, {"line": 38, "name": "run_tests", "filename": 3}, {"line": 1273, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3116, "name": "run", "filename": 4}, {"line": 3088, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 2983, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 395, "name": "test_flex_attention_caching", "filename": 1}, {"line": 380, "name": "fn", "filename": 1}, {"line": 1062, "name": "flex_attention", "filename": 10}, {"line": 1049, "name": "_flex_attention_hop_wrapper", "filename": 10}]}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.632000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "d442394c2c9a973ef4b64ef5deddb37f"} - { - "name": "_compile.compile_inner", - "ts": 1727975450632606.5, - "args": null, +V1206 15:24:45.403000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "d89a3716f835eda9884b375881c8c021"} + { + "name": "dynamo", + "ts": 1733527485402945.8, + "args": { + "compile_id": "0/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:50.633000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "686bcefec76c75096eb61274391c2867"} +V1206 15:24:45.403000 1667746 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", 0]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/test/inductor/test_codecache.py", 1]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_inductor/test_case.py", 2]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_dynamo/test_case.py", 3]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/testing/_internal/common_utils.py", 4]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/main.py", 5]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/runner.py", 6]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/suite.py", 7]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/case.py", 8]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/contextlib.py", 9]} +V1206 15:24:45.404000 1667746 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py", 10]} +V1206 15:24:45.405000 1667746 torch/_dynamo/convert_frame.py:961] {"dynamo_start": {"stack": [{"line": 1439, "name": "", "filename": 1}, {"line": 15, "name": "run_tests", "filename": 2}, {"line": 39, "name": "run_tests", "filename": 3}, {"line": 1353, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3234, "name": "run", "filename": 4}, {"line": 3206, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 3099, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 475, "name": "test_flex_attention_caching", "filename": 1}, {"line": 460, "name": "fn", "filename": 1}, {"line": 1316, "name": "flex_attention", "filename": 10}, {"line": 1303, "name": "_flex_attention_hop_wrapper", "filename": 10}]}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.405000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "e415c7d5114d113beb39e5b78fc4c96d"} { "name": "entire_frame_compile", - "ts": 1727975450632606.5, - "args": null, + "ts": 1733527485405153.2, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:50.639000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 0, "size": 524288}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.640000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.640000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 0, "source": "L['args'][0]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.835000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 1, "describer_id": 0, "size": 524288}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.836000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 4, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.836000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 4, "source": "L['args'][1]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.837000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 2, "describer_id": 0, "size": 524288}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.837000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 5, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.838000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 5, "source": "L['args'][2]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.839000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 3, "describer_id": 0, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.839000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 6, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.839000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 6, "source": "L['args'][4][0]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.840000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 4, "describer_id": 0, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.841000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.841000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 7, "source": "L['args'][4][1]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.842000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 5, "describer_id": 0, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.842000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.843000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 8, "source": "L['args'][4][2]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.844000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 6, "describer_id": 0, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.844000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.844000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 9, "source": "L['args'][4][3]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.845000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 7, "describer_id": 0, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.846000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.846000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 10, "source": "L['args'][4][4]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.847000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 8, "describer_id": 0, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.847000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.847000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 11, "source": "L['args'][4][5]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.848000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 9, "describer_id": 0, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.849000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 12, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.849000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 12, "source": "L['args'][4][6]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.850000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 10, "describer_id": 0, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.850000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 13, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.850000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 13, "source": "L['args'][4][7]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:50.864000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_output_graph": {"sizes": {"l_args_0_": [1, 4, 512, 64], "l_args_1_": [1, 4, 512, 64], "l_args_2_": [1, 4, 512, 64], "l_args_4_0_": [1, 1, 16], "l_args_4_1_": [1, 1, 16, 16], "l_args_4_2_": [1, 1, 16], "l_args_4_3_": [1, 1, 16, 16], "l_args_4_4_": [1, 1, 16], "l_args_4_5_": [1, 1, 16, 16], "l_args_4_6_": [1, 1, 16], "l_args_4_7_": [1, 1, 16, 16], "child_1": [], "child_2": [], "child_3": [], "child_4": [], "child": [], "child_5": [], "child_6": [], "child_7": [], "child_8": [], "getitem": [1, 4, 512, 64], "getitem_1": [1, 4, 512]}}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "36de5ad6eb1efc648a27dc62c107a2ca"} +V1206 15:24:45.409000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 0, "describer_id": 0, "size": 524288}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.409000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.409000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 0, "source": "L['args'][0]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.519000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 1, "describer_id": 0, "size": 524288}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.520000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 4, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.520000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 4, "source": "L['args'][1]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.520000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 2, "describer_id": 0, "size": 524288}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.521000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 5, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.521000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 5, "source": "L['args'][2]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.521000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 3, "describer_id": 0, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.521000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 6, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.522000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 6, "source": "L['args'][4][0]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.522000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 4, "describer_id": 0, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.522000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.522000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 7, "source": "L['args'][4][1]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.523000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 5, "describer_id": 0, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.523000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.523000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 8, "source": "L['args'][4][2]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.524000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 6, "describer_id": 0, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.524000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.524000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 9, "source": "L['args'][4][3]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.525000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 7, "describer_id": 0, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.525000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.525000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 10, "source": "L['args'][4][4]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.526000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 8, "describer_id": 0, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.526000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.526000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 11, "source": "L['args'][4][5]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.526000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 9, "describer_id": 0, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.526000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 12, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.527000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 12, "source": "L['args'][4][6]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.527000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 10, "describer_id": 0, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.527000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 13, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.527000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 13, "source": "L['args'][4][7]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.534000 1667746 torch/_dynamo/output_graph.py:1336] {"dynamo_output_graph": {"sizes": {"l_args_0_": [1, 4, 512, 64], "l_args_1_": [1, 4, 512, 64], "l_args_2_": [1, 4, 512, 64], "l_args_4_0_": [1, 1, 16], "l_args_4_1_": [1, 1, 16, 16], "l_args_4_2_": [1, 1, 16], "l_args_4_3_": [1, 1, 16, 16], "l_args_4_4_": [1, 1, 16], "l_args_4_5_": [1, 1, 16, 16], "l_args_4_6_": [1, 1, 16], "l_args_4_7_": [1, 1, 16, 16], "getitem": [1, 4, 512, 64], "getitem_1": [1, 4, 512]}}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "52fbb526c9a1f1575361cb8ed8e7794e"} class GraphModule(torch.nn.Module): def forward(self, L_args_0_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_args_1_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_args_2_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_args_4_0_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_1_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_args_4_2_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_3_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_args_4_4_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_5_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_args_4_6_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_7_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): l_args_0_ = L_args_0_ @@ -78,97 +83,62 @@ V1003 10:10:50.864000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_outpu l_args_4_6_ = L_args_4_6_ l_args_4_7_ = L_args_4_7_ - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1050 in _flex_attention_hop_wrapper, code: return flex_attention_hop(*args, **kwargs) - child_1: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_1 = None - child_2: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_2 = None - child_3: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_3 = None - child_4: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_4 = None - child: "f32[][]cuda:0" = l_args_0_.new_empty([], requires_grad = False); child = None + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1304 in _flex_attention_hop_wrapper, code: return flex_attention_hop(*args, **kwargs) score_mod_0 = self.score_mod_0 - child_5: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_5 = None - child_6: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_6 = None - child_7: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_7 = None - child_8: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_8 = None mask_fn_0 = self.mask_fn_0 - flex_attention = torch.ops.higher_order.flex_attention(l_args_0_, l_args_1_, l_args_2_, score_mod_0, (l_args_4_0_, l_args_4_1_, l_args_4_2_, l_args_4_3_, l_args_4_4_, l_args_4_5_, l_args_4_6_, l_args_4_7_, 128, 128, mask_fn_0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); l_args_0_ = l_args_1_ = l_args_2_ = score_mod_0 = l_args_4_0_ = l_args_4_1_ = l_args_4_2_ = l_args_4_3_ = l_args_4_4_ = l_args_4_5_ = l_args_4_6_ = l_args_4_7_ = mask_fn_0 = None + flex_attention = torch.ops.higher_order.flex_attention(l_args_0_, l_args_1_, l_args_2_, score_mod_0, (l_args_4_0_, l_args_4_1_, l_args_4_2_, l_args_4_3_, l_args_4_4_, l_args_4_5_, l_args_4_6_, l_args_4_7_, 128, 128, mask_fn_0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); l_args_0_ = l_args_1_ = l_args_2_ = score_mod_0 = l_args_4_0_ = l_args_4_1_ = l_args_4_2_ = l_args_4_3_ = l_args_4_4_ = l_args_4_5_ = l_args_4_6_ = l_args_4_7_ = mask_fn_0 = None getitem: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0] getitem_1: "f32[1, 4, 512][2048, 512, 1]cuda:0" = flex_attention[1]; flex_attention = None return (getitem, getitem_1) class score_mod_0(torch.nn.Module): def forward(self, child: "f32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0", child_4: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:377 in score_mod, code: return score + (q - kv) + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:457 in score_mod, code: return score + (q - kv) sub: "i32[][]cuda:0" = child_3 - child_4; child_3 = child_4 = None add: "f32[][]cuda:0" = child + sub; child = sub = None return add class mask_fn_0(torch.nn.Module): - def forward(self, child_5: "i32[][]cuda:0", child_6: "i32[][]cuda:0", child_7: "i32[][]cuda:0", child_8: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 - ge: "b8[][]cuda:0" = child_7 >= child_8; child_7 = child_8 = None + def forward(self, child: "i32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0"): + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + ge: "b8[][]cuda:0" = child_2 >= child_3; child_2 = child_3 = None return ge -V1003 10:10:50.865000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "78b3ca52339404a8a8b475af3393e9be"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975450865801.5, - "args": null, - "ph": "B", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0 - } -V1003 10:10:50.866000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "f7bc91b808bbb22050236e69d68d305e"} - { - "name": "backend_compile", - "ts": 1727975450865801.5, - "args": null, - "ph": "B", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0 - } -V1003 10:10:50.866000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "4a09b2e379da345bd3263251d88c5014"} +V1206 15:24:45.534000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "ca2a774ae001e87e57bd31e3fb982793"} { "name": "backend_compile", - "ts": 1727975450866663.5, + "ts": 1733527485534550.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 0, - "fxgraph_cache_bypass": 0 - } + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" }, - "ph": "E", + "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:50.867000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "53dff51299e2eac55b28bb862cc5cb64"} +V1206 15:24:45.534000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "dbd3ef3cad5d244dd2996256c488f716"} { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975450867057.0, + "name": "backend_compile", + "ts": 1733527485534930.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 0, - "fxgraph_cache_bypass": 0 - } + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:50.893000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_str": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "e693a6590122b61ee5c84da6a497ef23"} +V1206 15:24:45.547000 1667746 torch/_dynamo/guards.py:2315] {"dynamo_cpp_guards_str": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "c168856da88c1195cba9758754bcd1a1"} TREE_GUARD_MANAGER: +- RootGuardManager - | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:471 in init_ambient_guards + | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:484 in init_ambient_guards | +- GLOBAL_STATE: ___check_global_state() | +- TORCH_FUNCTION_MODE_STACK: ___check_torch_function_mode_stack() - | +- GuardManager: source=L['args'], accessed_by=DictGetItemGuardAccessor(args) - | | +- TYPE_MATCH: ___check_type_id(L['args'], 8815232) + | +- GuardManager: source=L['args'], accessed_by=DictGetItemGuardAccessor('args') + | | +- TYPE_MATCH: ___check_type_id(L['args'], 8812224) | | +- LENGTH_CHECK: len(L['args']) == 7 | | +- GuardManager: source=L['args'][0], accessed_by=TupleGetItemGuardAccessor(0) | | | +- TENSOR_MATCH: check_tensor(L['args'][0], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) @@ -184,9 +154,9 @@ V1003 10:10:50.893000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | +- NO_TENSOR_ALIASING | | +- GuardManager: source=L['args'][3], accessed_by=TupleGetItemGuardAccessor(3) | | | +- GuardManager: source=L['args'][3].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][3].__code__, 140413271879296) + | | | | +- ID_MATCH: ___check_obj_id(L['args'][3].__code__, 140062268556144) | | +- GuardManager: source=L['args'][4], accessed_by=TupleGetItemGuardAccessor(4) - | | | +- TYPE_MATCH: ___check_type_id(L['args'][4], 8815232) + | | | +- TYPE_MATCH: ___check_type_id(L['args'][4], 8812224) | | | +- LENGTH_CHECK: len(L['args'][4]) == 11 | | | +- GuardManager: source=L['args'][4][0], accessed_by=TupleGetItemGuardAccessor(0) | | | | +- TENSOR_MATCH: check_tensor(L['args'][4][0], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.int32, device=0, requires_grad=False, size=[1, 1, 16], stride=[16, 16, 1]) @@ -226,52 +196,60 @@ V1003 10:10:50.893000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | +- EQUALS_MATCH: L['args'][4][9] == 128 | | | +- GuardManager: source=L['args'][4][10], accessed_by=TupleGetItemGuardAccessor(10) | | | | +- GuardManager: source=L['args'][4][10].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(L['args'][4][10].__code__, 140413271880128) + | | | | | +- ID_MATCH: ___check_obj_id(L['args'][4][10].__code__, 140062268556768) | | +- GuardManager: source=L['args'][5], accessed_by=TupleGetItemGuardAccessor(5) | | | +- EQUALS_MATCH: L['args'][5] == 0.125 | | +- GuardManager: source=L['args'][6], accessed_by=TupleGetItemGuardAccessor(6) - | | | +- DICT_LENGTH: len(L['args'][6]) == 3 - | | | +- GuardManager: source=L['args'][6]['ROWS_GUARANTEED_SAFE'], accessed_by=DictGetItemGuardAccessor(ROWS_GUARANTEED_SAFE) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['ROWS_GUARANTEED_SAFE'], 8910592) - | | | +- GuardManager: source=L['args'][6]['PRESCALE_QK'], accessed_by=DictGetItemGuardAccessor(PRESCALE_QK) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['PRESCALE_QK'], 8910592) - | | | +- GuardManager: source=L['args'][6]['OUTPUT_LOGSUMEXP'], accessed_by=DictGetItemGuardAccessor(OUTPUT_LOGSUMEXP) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['OUTPUT_LOGSUMEXP'], 8910592) - | +- GuardManager: source=L['kwargs'], accessed_by=DictGetItemGuardAccessor(kwargs) + | | | +- DICT_LENGTH: len(L['args'][6]) == 4 + | | | +- GuardManager: source=L['args'][6]['PRESCALE_QK'], accessed_by=DictGetItemGuardAccessor('PRESCALE_QK') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['PRESCALE_QK'], 8907584) + | | | +- GuardManager: source=L['args'][6]['ROWS_GUARANTEED_SAFE'], accessed_by=DictGetItemGuardAccessor('ROWS_GUARANTEED_SAFE') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['ROWS_GUARANTEED_SAFE'], 8907584) + | | | +- GuardManager: source=L['args'][6]['BLOCKS_ARE_CONTIGUOUS'], accessed_by=DictGetItemGuardAccessor('BLOCKS_ARE_CONTIGUOUS') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['BLOCKS_ARE_CONTIGUOUS'], 8907584) + | | | +- GuardManager: source=L['args'][6]['OUTPUT_LOGSUMEXP'], accessed_by=DictGetItemGuardAccessor('OUTPUT_LOGSUMEXP') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['OUTPUT_LOGSUMEXP'], 8908032) + | +- GuardManager: source=L['kwargs'], accessed_by=DictGetItemGuardAccessor('kwargs') | | +- DICT_LENGTH: not L['kwargs'] | +- GuardManager: source=G, accessed_by=GlobalsGuardAccessor - | | +- GuardManager: source=G['flex_attention_hop'], accessed_by=DictGetItemGuardAccessor(flex_attention_hop) - | | | +- TYPE_MATCH: ___check_type_id(G['flex_attention_hop'], 96992544) + | | +- GuardManager: source=G['flex_attention_hop'], accessed_by=DictGetItemGuardAccessor('flex_attention_hop') + | | | +- TYPE_MATCH: ___check_type_id(G['flex_attention_hop'], 96230624) | | | +- GuardManager: source=G['flex_attention_hop'].__name__, accessed_by=GetAttrGuardAccessor(__name__) | | | | +- EQUALS_MATCH: G['flex_attention_hop'].__name__ == 'flex_attention' - | | +- GuardManager: source=G['__builtins_dict___0'], accessed_by=DictGetItemGuardAccessor(__builtins_dict___0) - | | | +- GuardManager: source=G['__builtins_dict___0']['len'], accessed_by=DictGetItemGuardAccessor(len) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['len'], 140413275558816) - | | | +- GuardManager: source=G['__builtins_dict___0']['sum'], accessed_by=DictGetItemGuardAccessor(sum) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['sum'], 140413275559936) - | | | +- GuardManager: source=G['__builtins_dict___0']['list'], accessed_by=DictGetItemGuardAccessor(list) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['list'], 8844320) - | | | +- GuardManager: source=G['__builtins_dict___0']['type'], accessed_by=DictGetItemGuardAccessor(type) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['type'], 8813248) - | | | +- GuardManager: source=G['__builtins_dict___0']['tuple'], accessed_by=DictGetItemGuardAccessor(tuple) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['tuple'], 8815232) - | | | +- GuardManager: source=G['__builtins_dict___0']['object'], accessed_by=DictGetItemGuardAccessor(object) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['object'], 8813984) - | | | +- GuardManager: source=G['__builtins_dict___0']['isinstance'], accessed_by=DictGetItemGuardAccessor(isinstance) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['isinstance'], 140413275558496) - | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_utils_dot__pytree) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140411217627952) + | | +- GuardManager: source=G['_140062263790704_c0'], accessed_by=DictGetItemGuardAccessor('_140062263790704_c0') + | | | +- GuardManager: source=G['_140062263790704_c0'].Tensor, accessed_by=GetAttrGuardAccessor(Tensor) + | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c0'].Tensor, 82181376) + | | | | +- GuardManager: source=G['_140062263790704_c0'].Tensor.__bases__, accessed_by=GetAttrGuardAccessor(__bases__) + | | | | | +- GuardManager: source=G['_140062263790704_c0'].Tensor.__bases__[0], accessed_by=TupleGetItemGuardAccessor(0) + | | | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c0'].Tensor.__bases__[0], 140062119703136) + | | +- GuardManager: source=G['__builtins_dict___0'], accessed_by=DictGetItemGuardAccessor('__builtins_dict___0') + | | | +- GuardManager: source=G['__builtins_dict___0']['len'], accessed_by=DictGetItemGuardAccessor('len') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['len'], 140062269592480) + | | | +- GuardManager: source=G['__builtins_dict___0']['sum'], accessed_by=DictGetItemGuardAccessor('sum') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['sum'], 140062269593600) + | | | +- GuardManager: source=G['__builtins_dict___0']['list'], accessed_by=DictGetItemGuardAccessor('list') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['list'], 8841312) + | | | +- GuardManager: source=G['__builtins_dict___0']['type'], accessed_by=DictGetItemGuardAccessor('type') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['type'], 8810240) + | | | +- GuardManager: source=G['__builtins_dict___0']['tuple'], accessed_by=DictGetItemGuardAccessor('tuple') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['tuple'], 8812224) + | | | +- GuardManager: source=G['__builtins_dict___0']['object'], accessed_by=DictGetItemGuardAccessor('object') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['object'], 8810976) + | | | +- GuardManager: source=G['__builtins_dict___0']['isinstance'], accessed_by=DictGetItemGuardAccessor('isinstance') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['isinstance'], 140062269592160) + | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_utils_dot__pytree') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140057600329744) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].TreeSpec, accessed_by=GetAttrGuardAccessor(TreeSpec) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 84866496) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 86791872) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf, accessed_by=GetAttrGuardAccessor(_is_leaf) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140411217262720) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140057601749808) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, accessed_by=GetAttrGuardAccessor(_LEAF_SPEC) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 85171104) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 87025712) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, accessed_by=GetAttrGuardAccessor(type) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, accessed_by=GetAttrGuardAccessor(context) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes, accessed_by=GetAttrGuardAccessor(num_nodes) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes == 1 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_leaves, accessed_by=GetAttrGuardAccessor(num_leaves) @@ -279,116 +257,135 @@ V1003 10:10:50.893000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children, accessed_by=GetAttrGuardAccessor(num_children) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children == 0 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, accessed_by=GetAttrGuardAccessor(children_specs) - | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8844320) + | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8841312) | | | | | +- LENGTH_CHECK: not G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type, accessed_by=GetAttrGuardAccessor(_get_node_type) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140411217262448) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140057601749536) | | | +- DictGuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES, accessed_by=GetAttrGuardAccessor(SUPPORTED_NODES) - | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 519596 + | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 366913 | | | | +- KeyValueManager pair at index=1 | | | | | +- ValueManager: GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]] | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn, accessed_by=GetAttrGuardAccessor(flatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140411196281984) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140057599989696) | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn, accessed_by=GetAttrGuardAccessor(unflatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140411217182288) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140057600267792) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper, accessed_by=GetAttrGuardAccessor(_tree_flatten_helper) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140411217413040) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140057600001392) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance, accessed_by=GetAttrGuardAccessor(_is_namedtuple_instance) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140411217412592) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140057600000944) -V1003 10:10:50.894000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "20353735c197adbac78fed570e5a0d91"} +V1206 15:24:45.547000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "9f584d7434070bbe2c589c1cc5c9fa1c"} { "name": "entire_frame_compile", - "ts": 1727975450894280.0, + "ts": 1733527485547627.0, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 0, - "fxgraph_cache_bypass": 0 - } + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:50.894000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "6dc3af6842c03572276aaba0318e48d3"} +V1206 15:24:45.550000 1667746 torch/_dynamo/utils.py:1045] {"compilation_metrics": {"aot_autograd_cumulative_compile_time_us": 379, "dynamo_cumulative_compile_time_us": 142473, "frame_key": "1", "co_name": "_flex_attention_hop_wrapper", "co_filename": "/data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py", "co_firstlineno": 1303, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 58, "shape_env_guard_count": 0, "graph_op_count": 3, "graph_node_count": 17, "graph_input_count": 11, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": false, "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 9223372036854775807, \"accumulated_cache_size_limit\": 9223372036854775807, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}", "is_forward": true, "dynamo_compile_time_before_restart_us": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.550000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "4808e1bc6347d31bf5778999c536de01"} { - "name": "_compile.compile_inner", - "ts": 1727975450894591.0, + "name": "dynamo", + "ts": 1733527485550622.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 0, - "fxgraph_cache_bypass": 0 - } + "compile_id": "0/0", + "frame_key": "1", + "co_name": "_flex_attention_hop_wrapper", + "co_filename": "/data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py", + "co_firstlineno": 1303, + "cache_size": 0, + "accumulated_cache_size": 0, + "guard_count": 58, + "shape_env_guard_count": 0, + "graph_op_count": 3, + "graph_node_count": 17, + "graph_input_count": 11, + "fail_type": null, + "fail_reason": null, + "fail_user_frame_filename": null, + "fail_user_frame_lineno": null, + "non_compliant_ops": [], + "compliant_custom_ops": [], + "restart_reasons": [], + "dynamo_time_before_restart_s": 0.0, + "has_guarded_code": true, + "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 9223372036854775807, \"accumulated_cache_size_limit\": 9223372036854775807, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:50.895000 2235078 torch/_dynamo/utils.py:840] {"compilation_metrics": {"compile_id": "0/0", "frame_key": "1", "co_name": "_flex_attention_hop_wrapper", "co_filename": "/data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py", "co_firstlineno": 1049, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 55, "shape_env_guard_count": 0, "graph_op_count": 12, "graph_node_count": 26, "graph_input_count": 11, "start_time": 1727975450.6325822, "entire_frame_compile_time_s": 0.26139068603515625, "backend_compile_time_s": 0.0006465911865234375, "inductor_compile_time_s": null, "code_gen_time_s": null, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "possibly_missed_reinplacing_opportunities": 0, "remote_cache_time_saved_s": 0, "structured_logging_overhead_s": 0.019217516000000004, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": true}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.499000 2235078 torch/_dynamo/convert_frame.py:915] {"dynamo_start": {"stack": [{"line": 916, "name": "", "filename": 1}, {"line": 14, "name": "run_tests", "filename": 2}, {"line": 38, "name": "run_tests", "filename": 3}, {"line": 1273, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3116, "name": "run", "filename": 4}, {"line": 3088, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 2983, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 395, "name": "test_flex_attention_caching", "filename": 1}, {"line": 379, "name": "fn", "filename": 1}]}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.499000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "e10001bc2249e6200cda3bf42b80be57"} +V1206 15:24:45.976000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b144b519a5591ad277125a12c84b389b"} { - "name": "_compile.compile_inner", - "ts": 1727975453499593.8, - "args": null, + "name": "dynamo", + "ts": 1733527485976303.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:53.499000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ff0561bc609e54d7e5775bdf8f81ff48"} +V1206 15:24:45.976000 1667746 torch/_dynamo/convert_frame.py:961] {"dynamo_start": {"stack": [{"line": 1439, "name": "", "filename": 1}, {"line": 15, "name": "run_tests", "filename": 2}, {"line": 39, "name": "run_tests", "filename": 3}, {"line": 1353, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3234, "name": "run", "filename": 4}, {"line": 3206, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 3099, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 475, "name": "test_flex_attention_caching", "filename": 1}, {"line": 459, "name": "fn", "filename": 1}]}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.977000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b2b1c355b023de0af140013ba14024ed"} { "name": "entire_frame_compile", - "ts": 1727975453499593.8, - "args": null, + "ts": 1733527485977092.0, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:53.502000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 9, "size": 524288}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.503000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.503000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 0, "source": "L['q']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.511000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 1, "describer_id": 9, "size": 524288}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.512000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 1, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.512000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 1, "source": "L['k']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.513000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 2, "describer_id": 9, "size": 524288}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.514000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 2, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.514000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 2, "source": "L['v']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.532000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 3, "describer_id": 9, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.532000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 3, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.532000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 3, "source": "L['block_mask'].kv_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.676000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 4, "describer_id": 9, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.677000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.677000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 7, "source": "L['block_mask'].kv_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.679000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 5, "describer_id": 9, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.679000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.679000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 8, "source": "L['block_mask'].full_kv_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.680000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 6, "describer_id": 9, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.681000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.681000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 9, "source": "L['block_mask'].full_kv_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.682000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 7, "describer_id": 9, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.683000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.683000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 10, "source": "L['block_mask'].q_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.684000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 8, "describer_id": 9, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.685000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.685000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 11, "source": "L['block_mask'].q_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.686000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 9, "describer_id": 9, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.686000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 12, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.687000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 12, "source": "L['block_mask'].full_q_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.688000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 10, "describer_id": 9, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.688000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 13, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 9}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.689000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 9, "id": 13, "source": "L['block_mask'].full_q_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:10:53.697000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_output_graph": {"sizes": {"l_q_": [1, 4, 512, 64], "l_k_": [1, 4, 512, 64], "l_v_": [1, 4, 512, 64], "l_block_mask_kv_num_blocks": [1, 1, 16], "l_block_mask_kv_indices": [1, 1, 16, 16], "l_block_mask_full_kv_num_blocks": [1, 1, 16], "l_block_mask_full_kv_indices": [1, 1, 16, 16], "l_block_mask_q_num_blocks": [1, 1, 16], "l_block_mask_q_indices": [1, 1, 16, 16], "l_block_mask_full_q_num_blocks": [1, 1, 16], "l_block_mask_full_q_indices": [1, 1, 16, 16], "child_1": [], "child_2": [], "child_3": [], "child_4": [], "child": [], "child_5": [], "child_6": [], "child_7": [], "child_8": [], "out": [1, 4, 512, 64]}}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0365bd76b3474fb6d96e4c3c42585fb7"} +V1206 15:24:45.979000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 0, "describer_id": 9, "size": 524288}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.979000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.979000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 0, "source": "L['q']"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.985000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 1, "describer_id": 9, "size": 524288}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.986000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 1, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.986000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 1, "source": "L['k']"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.987000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 2, "describer_id": 9, "size": 524288}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.987000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 2, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.987000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 2, "source": "L['v']"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.998000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 3, "describer_id": 9, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.998000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 3, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:45.998000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 3, "source": "L['block_mask'].kv_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.071000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 4, "describer_id": 9, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.071000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.071000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 7, "source": "L['block_mask'].kv_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.072000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 5, "describer_id": 9, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.072000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.072000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 8, "source": "L['block_mask'].full_kv_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.073000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 6, "describer_id": 9, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.073000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.073000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 9, "source": "L['block_mask'].full_kv_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.074000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 7, "describer_id": 9, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.074000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.074000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 10, "source": "L['block_mask'].q_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.075000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 8, "describer_id": 9, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.075000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.075000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 11, "source": "L['block_mask'].q_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.076000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 9, "describer_id": 9, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.076000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 12, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.076000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 12, "source": "L['block_mask'].full_q_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.077000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 10, "describer_id": 9, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.077000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 13, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 9}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.077000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 9, "id": 13, "source": "L['block_mask'].full_q_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:46.082000 1667746 torch/_dynamo/output_graph.py:1336] {"dynamo_output_graph": {"sizes": {"l_q_": [1, 4, 512, 64], "l_k_": [1, 4, 512, 64], "l_v_": [1, 4, 512, 64], "l_block_mask_kv_num_blocks": [1, 1, 16], "l_block_mask_kv_indices": [1, 1, 16, 16], "l_block_mask_full_kv_num_blocks": [1, 1, 16], "l_block_mask_full_kv_indices": [1, 1, 16, 16], "l_block_mask_q_num_blocks": [1, 1, 16], "l_block_mask_q_indices": [1, 1, 16, 16], "l_block_mask_full_q_num_blocks": [1, 1, 16], "l_block_mask_full_q_indices": [1, 1, 16, 16], "out": [1, 4, 512, 64]}}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "e6aa2ecb5d533ab181a6215ce16f359c"} class GraphModule(torch.nn.Module): def forward(self, L_q_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_k_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_v_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_block_mask_kv_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_kv_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_block_mask_full_kv_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_full_kv_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_block_mask_q_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_q_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_block_mask_full_q_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_full_q_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): l_q_ = L_q_ @@ -403,119 +400,316 @@ V1003 10:10:53.697000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_outpu l_block_mask_full_q_num_blocks = L_block_mask_full_q_num_blocks l_block_mask_full_q_indices = L_block_mask_full_q_indices - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( - child_1: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_1 = None - child_2: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_2 = None - child_3: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_3 = None - child_4: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_4 = None - child: "f32[][]cuda:0" = l_q_.new_empty([], requires_grad = False); child = None + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( score_mod_0 = self.score_mod_0 - child_5: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_5 = None - child_6: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_6 = None - child_7: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_7 = None - child_8: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_8 = None mask_fn_0 = self.mask_fn_0 - flex_attention = torch.ops.higher_order.flex_attention(l_q_, l_k_, l_v_, score_mod_0, (l_block_mask_kv_num_blocks, l_block_mask_kv_indices, l_block_mask_full_kv_num_blocks, l_block_mask_full_kv_indices, l_block_mask_q_num_blocks, l_block_mask_q_indices, l_block_mask_full_q_num_blocks, l_block_mask_full_q_indices, 128, 128, mask_fn_0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); l_q_ = l_k_ = l_v_ = score_mod_0 = l_block_mask_kv_num_blocks = l_block_mask_kv_indices = l_block_mask_full_kv_num_blocks = l_block_mask_full_kv_indices = l_block_mask_q_num_blocks = l_block_mask_q_indices = l_block_mask_full_q_num_blocks = l_block_mask_full_q_indices = mask_fn_0 = None + flex_attention = torch.ops.higher_order.flex_attention(l_q_, l_k_, l_v_, score_mod_0, (l_block_mask_kv_num_blocks, l_block_mask_kv_indices, l_block_mask_full_kv_num_blocks, l_block_mask_full_kv_indices, l_block_mask_q_num_blocks, l_block_mask_q_indices, l_block_mask_full_q_num_blocks, l_block_mask_full_q_indices, 128, 128, mask_fn_0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); l_q_ = l_k_ = l_v_ = score_mod_0 = l_block_mask_kv_num_blocks = l_block_mask_kv_indices = l_block_mask_full_kv_num_blocks = l_block_mask_full_kv_indices = l_block_mask_q_num_blocks = l_block_mask_q_indices = l_block_mask_full_q_num_blocks = l_block_mask_full_q_indices = mask_fn_0 = None out: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0]; flex_attention = None return (out,) class score_mod_0(torch.nn.Module): def forward(self, child: "f32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0", child_4: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:377 in score_mod, code: return score + (q - kv) + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:457 in score_mod, code: return score + (q - kv) sub: "i32[][]cuda:0" = child_3 - child_4; child_3 = child_4 = None add: "f32[][]cuda:0" = child + sub; child = sub = None return add class mask_fn_0(torch.nn.Module): - def forward(self, child_5: "i32[][]cuda:0", child_6: "i32[][]cuda:0", child_7: "i32[][]cuda:0", child_8: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 - ge: "b8[][]cuda:0" = child_7 >= child_8; child_7 = child_8 = None + def forward(self, child: "i32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0"): + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + ge: "b8[][]cuda:0" = child_2 >= child_3; child_2 = child_3 = None return ge -V1003 10:10:53.698000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f3daea7b3445ed3e94cb46786b2c5d08"} +V1206 15:24:46.082000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "37b24b3f85c3e5d89e19d2ed089a024a"} { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975453698084.0, - "args": null, + "name": "backend_compile", + "ts": 1733527486082618.8, + "args": { + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:53.698000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d8ed1bc2cb1632abc21c22f112ed0fec"} +V1206 15:24:46.082000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "1cb1b51c68c2aa91138743ed176863dc"} { - "name": "backend_compile", - "ts": 1727975453698084.0, - "args": null, + "name": "_recursive_pre_grad_passes", + "ts": 1733527486082898.8, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:53.708000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "18faf90ed7d7bb7f081083da150205b3"} +V1206 15:24:46.085000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "fe96e7264be7dc8710649f24088a12a7"} + { + "name": "_recursive_pre_grad_passes", + "ts": 1733527486085403.5, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:46.087000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f3b0a502ac862b43ef8e578cbbbedaef"} { "name": "create_aot_dispatcher_function", - "ts": 1727975453708777.0, - "args": null, + "ts": 1733527486087844.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:53.842000 2235078 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:215] {"aot_forward_graph": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f6cec159cb1cee52405afcfe40b62c03"} +V1206 15:24:46.137000 1667746 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:214] {"artifact": {"name": "aot_forward_graph_fw_metadata", "encoding": "string"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d952b3071dd9ba606a04d644841d9aba"} + ViewAndMutationMeta(input_info=[InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True)], + output_info=[OutputAliasInfo(output_type=, + raw_type=, + base_idx=None, + dynamic_dims=set(), + requires_grad=False, + functional_tensor=None)], + num_intermediate_bases=0, + keep_input_mutations=True, + traced_tangents=[], + subclass_inp_meta=[PlainTensorMeta(unwrapped_idx=0, + memory_format=None), + PlainTensorMeta(unwrapped_idx=1, + memory_format=None), + PlainTensorMeta(unwrapped_idx=2, + memory_format=None), + PlainTensorMeta(unwrapped_idx=3, + memory_format=None), + PlainTensorMeta(unwrapped_idx=4, + memory_format=None), + PlainTensorMeta(unwrapped_idx=5, + memory_format=None), + PlainTensorMeta(unwrapped_idx=6, + memory_format=None), + PlainTensorMeta(unwrapped_idx=7, + memory_format=None), + PlainTensorMeta(unwrapped_idx=8, + memory_format=None), + PlainTensorMeta(unwrapped_idx=9, + memory_format=None), + PlainTensorMeta(unwrapped_idx=10, + memory_format=None)], + subclass_fw_graph_out_meta=[PlainTensorMeta(unwrapped_idx=0, + memory_format=None)], + subclass_tangent_meta=[], + is_train=False, + traced_tangent_metas=None, + num_symints_saved_for_bw=None, + grad_enabled_mutation=None, + deterministic=None, + static_input_indices=[], + tokens={}, + indices_of_inputs_that_requires_grad_with_mutations_in_bw=[], + bw_donated_idxs=None, + num_backward_tokens=0) +V1206 15:24:46.138000 1667746 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:232] {"aot_inference_graph": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "02884732a6f03bac463b0e1e5f9b1cf4"} class (torch.nn.Module): def forward(self, arg0_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg1_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg2_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg3_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg4_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg5_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg6_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg7_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg8_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg9_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg10_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( sdpa_score0 = self.sdpa_score0 sdpa_mask0 = self.sdpa_mask0 - flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None + flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None getitem: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0]; flex_attention = None return (getitem,) class sdpa_score0(torch.nn.Module): def forward(self, arg0_1: "f32[][]cuda:0", arg1_1: "i32[][]cuda:0", arg2_1: "i32[][]cuda:0", arg3_1: "i32[][]cuda:0", arg4_1: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( sub: "i32[][]cuda:0" = torch.ops.aten.sub.Tensor(arg3_1, arg4_1); arg3_1 = arg4_1 = None add: "f32[][]cuda:0" = torch.ops.aten.add.Tensor(arg0_1, sub); arg0_1 = sub = None return add class sdpa_mask0(torch.nn.Module): def forward(self, arg0_1: "i32[][]cuda:0", arg1_1: "i32[][]cuda:0", arg2_1: "i32[][]cuda:0", arg3_1: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 ge: "b8[][]cuda:0" = torch.ops.aten.ge.Tensor(arg2_1, arg3_1); arg2_1 = arg3_1 = None return ge -V1003 10:10:53.843000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "5dd81e8de7669fa4f8da5f234591e1dc"} +V1206 15:24:46.139000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "41345436bd100a92ac2e1c11dbcfde73"} { "name": "compile_fx..fw_compiler_base", - "ts": 1727975453842696.2, - "args": null, + "ts": 1733527486139665.0, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:54.072000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "df64f9080a1fb9082793b2c995d4ba13"} +V1206 15:24:46.140000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ff1f544bdf98bb91144ed9d00ccfd2c4"} { - "name": "compile_fx_inner", - "ts": 1727975454072213.5, - "args": null, + "name": "_recursive_joint_graph_passes", + "ts": 1733527486140025.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:54.072000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "6fe6927d5157dc6d0578283e86ead8a3"} +V1206 15:24:46.240000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b7986424633a93e63eaef105f6813984"} + { + "name": "_recursive_joint_graph_passes", + "ts": 1733527486240539.8, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:46.241000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "71eb226a27256d687d98919bf6379b6e"} { "name": "inductor_compile", - "ts": 1727975454072213.5, - "args": null, + "ts": 1733527486240972.5, + "args": { + "fn_name": "compile_fx_inner", + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:46.241000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "e3aa25638cfc1bbeacc1a3b1146ee001"} + { + "name": "inductor_codecache_torch_key", + "ts": 1733527486241893.2, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.234000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {"name": "fx_graph_runnable", "encoding": "string"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c507028eda974d39e4ca6593c6320460"} +V1206 15:24:46.497000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "2357634e283643a8bf96005629e2a966"} + { + "name": "inductor_codecache_torch_key", + "ts": 1733527486497765.5, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:46.953000 1667746 torch/_inductor/compile_fx.py:835] {"artifact": {"name": "fx_graph_runnable", "encoding": "string"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "2685e2fe8ae1c10315e3657921ebac7a"} import torch from torch import tensor, device @@ -528,11 +722,40 @@ V1003 10:10:56.234000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {" import torch._inductor.config import torch._functorch.config import torch.fx.experimental._config + torch._dynamo.config.cache_size_limit = 8 + torch._dynamo.config.accumulated_cache_size_limit = 256 + torch._dynamo.config.traceable_tensor_subclasses = set() + torch._dynamo.config.suppress_errors = False + torch._dynamo.config.allowed_functions_module_string_ignorelist = {'torch._decomp', 'torch.testing', 'torch._prims', 'torch._refs', 'torch.distributions'} + torch._dynamo.config._ddp_optimization_mode = ['ddp_optimizer', 'python_reducer', 'python_reducer_without_compiled_forward', 'no_optimization'] + torch._dynamo.config.raise_on_ctx_manager_usage = True + torch._dynamo.config._save_config_ignore = {'repro_after', 'skipfiles_inline_module_allowlist', 'repro_level', 'constant_functions'} torch._dynamo.config.log_compilation_metrics = False + torch._dynamo.config.reorderable_logging_functions = set() + torch._dynamo.config._autograd_backward_strict_mode_banned_ops = ['stride', 'requires_grad', 'storage_offset', 'layout', 'data', 'is_coalesced', 'is_complex', 'is_conj', 'is_contiguous', 'is_cpu', 'is_cuda', 'is_distributed', 'is_floating_point', 'is_inference', 'is_ipu', 'is_leaf', 'is_maia', 'is_meta', 'is_mkldnn', 'is_mps', 'is_mtia', 'is_neg', 'is_nested', 'is_nonzero', 'is_pinned', 'is_quantized', 'is_same_size', 'is_set_to', 'is_shared', 'is_signed', 'is_sparse', 'is_sparse_csr', 'is_vulkan', 'is_xla', 'is_xpu'] + torch._dynamo.config.fake_tensor_cache_enabled = True torch._dynamo.config.fake_tensor_cache_crosscheck_enabled = True + torch._dynamo.config.compiled_autograd_kwargs_override = {} + torch._inductor.config.fx_graph_cache = True torch._inductor.config.fx_graph_remote_cache = False torch._inductor.config.autotune_local_cache = False torch._inductor.config.autotune_remote_cache = False + torch._inductor.config.bundled_autotune_remote_cache = False + torch._inductor.config.pre_grad_fusion_options = {} + torch._inductor.config.post_grad_fusion_options = {} + torch._inductor.config.fx_passes_numeric_check = {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True} + torch._inductor.config.reorder_for_compute_comm_overlap_passes = ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms'] + torch._inductor.config._fuse_ddp_communication_passes = ['fuse_ddp_with_concat_op', 'schedule_comm_wait'] + torch._inductor.config.aot_inductor.metadata = {} + torch._inductor.config.aot_inductor.presets = {} + torch._inductor.config.rocm.arch = [] + torch._inductor.config.rocm.ck_supported_arch = ['gfx90a', 'gfx940', 'gfx941', 'gfx942'] + torch._inductor.config._save_config_ignore = ['trace.upload_tar', 'joint_custom_pre_pass', 'joint_custom_post_pass', 'pre_grad_custom_pass'] + torch._inductor.config._cache_config_ignore_prefix = ['trace', 'cuda.cutlass_dir', 'worker_start_method', 'compile_threads', 'post_grad_custom_post_pass', 'post_grad_custom_pre_pass', 'always_complex_memory_overlap_TESTING_ONLY'] + torch._inductor.config.external_matmul = [] + torch._functorch.config.functionalize_rng_ops = False + torch._functorch.config.enable_autograd_cache = True + torch._functorch.config.fake_tensor_allow_unsafe_data_ptr_access = True torch._functorch.config.unlift_effect_tokens = True @@ -541,20 +764,21 @@ V1003 10:10:56.234000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {" - # torch version: 2.5.0a0+git7647c39 - # torch cuda version: 12.0 - # torch git version: 7647c398ff87daf70260854cf0a7f7993b3abc76 + + # torch version: 2.6.0a0+giteece9ec + # torch cuda version: 12.2 + # torch git version: eece9ecd62cae84bc2f915fc48cffe43e30256aa # CUDA Info: # nvcc: NVIDIA (R) Cuda compiler driver # Copyright (c) 2005-2023 NVIDIA Corporation - # Built on Fri_Jan__6_16:45:21_PST_2023 - # Cuda compilation tools, release 12.0, V12.0.140 - # Build cuda_12.0.r12.0/compiler.32267302_0 + # Built on Tue_Aug_15_22:02:13_PDT_2023 + # Cuda compilation tools, release 12.2, V12.2.140 + # Build cuda_12.2.r12.2/compiler.33191640_0 # GPU Hardware Info: - # NVIDIA PG509-210 : 8 + # NVIDIA H100 : 8 from torch.nn import * @@ -569,7 +793,7 @@ V1003 10:10:56.234000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {" def forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1): sdpa_score0 = self.sdpa_score0 sdpa_mask0 = self.sdpa_mask0 - flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None + flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None getitem = flex_attention[0]; flex_attention = None return (getitem,) @@ -605,154 +829,223 @@ V1003 10:10:56.234000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {" # To run it separately, do # mod, args = run_repro(mod, load_args, accuracy=False, command='get_args', save_dir=None, tracing_mode='real', check_str=None) # mod(*args) -V1003 10:10:56.463000 2235078 torch/_inductor/compile_fx.py:795] {"inductor_post_grad_graph": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f6cec159cb1cee52405afcfe40b62c03"} +V1206 15:24:46.957000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "59eead8c2c99ba4c19b6f2e1672cfb1c"} + { + "name": "_recursive_post_grad_passes", + "ts": 1733527486957768.2, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:46.966000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "65cd5e088aa4752db9445bb36e57e3aa"} + { + "name": "_recursive_post_grad_passes", + "ts": 1733527486966834.5, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:46.968000 1667746 torch/_inductor/compile_fx.py:898] {"inductor_post_grad_graph": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "02884732a6f03bac463b0e1e5f9b1cf4"} class (torch.nn.Module): def forward(self, arg0_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg1_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg2_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg3_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg4_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg5_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg6_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg7_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg8_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg9_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg10_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( sdpa_score0 = self.sdpa_score0 sdpa_mask0 = self.sdpa_mask0 - flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None + flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None getitem: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0]; flex_attention = None return (getitem,) class sdpa_score0(torch.nn.Module): def forward(self, arg0_1: "f32[][]cuda:0", arg1_1: "i32[][]cuda:0", arg2_1: "i32[][]cuda:0", arg3_1: "i32[][]cuda:0", arg4_1: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( sub: "i32[][]cuda:0" = torch.ops.aten.sub.Tensor(arg3_1, arg4_1); arg3_1 = arg4_1 = None add: "f32[][]cuda:0" = torch.ops.aten.add.Tensor(arg0_1, sub); arg0_1 = sub = None return add class sdpa_mask0(torch.nn.Module): def forward(self, arg0_1: "i32[][]cuda:0", arg1_1: "i32[][]cuda:0", arg2_1: "i32[][]cuda:0", arg3_1: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 ge: "b8[][]cuda:0" = torch.ops.aten.ge.Tensor(arg2_1, arg3_1); arg2_1 = arg3_1 = None return ge -V1003 10:10:56.496000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "fc5eec6dd74463eae278f61fbe01c218"} +V1206 15:24:46.972000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "48d0779166f135a77ff5cde83cf71d3e"} { "name": "GraphLowering.run", - "ts": 1727975456496145.2, - "args": null, + "ts": 1733527486972460.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.814000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ceafe5eb3d5703b5ca9729fc880cd47b"} +V1206 15:24:47.138000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "6627ba3ae1644229670ab4d3fb0d4169"} { "name": "GraphLowering.run", - "ts": 1727975456813991.2, + "ts": 1733527487138160.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.814000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "5103da0c4e3f17344951cb8876a0ef08"} +V1206 15:24:47.138000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "1cd377482ee4a0ca5764fecedf20d954"} { - "name": "GraphLowering.compile_to_module", - "ts": 1727975456814764.0, - "args": null, + "name": "GraphLowering.compile_to_fn", + "ts": 1733527487138712.0, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.815000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0cc1abc97b9fd26bc3c8bcfa0fb94399"} +V1206 15:24:47.139000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f4590e34449b7bb09aed2c93c625760d"} { "name": "code_gen", - "ts": 1727975456814764.0, - "args": null, + "ts": 1733527487139024.8, + "args": { + "fn_name": "GraphLowering.compile_to_module", + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.820000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b680f8a68ec3cd182eef8fd76116df1f"} +V1206 15:24:47.139000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "7b65eb21df5478dbf8fba3b59850be45"} { - "name": "Scheduler.__init__", - "ts": 1727975456820242.5, - "args": null, + "name": "GraphLowering.codegen", + "ts": 1733527487139268.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.826000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "3ea7d3d1a284720131038fa646c6b219"} +V1206 15:24:47.140000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "9f9bd4fb5f640d8a4af811e82da8f2b7"} { "name": "Scheduler.__init__", - "ts": 1727975456826429.2, + "ts": 1733527487140643.2, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:47.144000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "36fb5520981a68e9cd0a7455ec7e42d1"} + { + "name": "Scheduler.fused_nodes", + "ts": 1733527487144035.5, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:47.144000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ff4c0585a764bffbe15620787ef7984e"} + { + "name": "Scheduler.fused_nodes", + "ts": 1733527487144372.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 } +V1206 15:24:47.146000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b2127891028545727d5538178e766dfd"} + { + "name": "Scheduler.__init__", + "ts": 1733527487146903.5, + "args": { + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.826000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "6328137516750744d43158caca6502ff"} +V1206 15:24:47.147000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "7d89af7c7d58849f171d9e78f71d86a2"} { "name": "Scheduler.codegen", - "ts": 1727975456826813.2, - "args": null, + "ts": 1733527487147125.2, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.837000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ec3d70ea4b3cc1330ead5e46c08a16ac"} +V1206 15:24:47.153000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "3b4150db66839c4586992ffee613eae4"} { "name": "Scheduler.codegen", - "ts": 1727975456837107.8, + "ts": 1733527487153011.0, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.837000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "bdcac968c06e063fd73d86a38c67b967"} +V1206 15:24:47.153000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "bf81162402f2ae6984f652a3f44baa84"} { "name": "PythonWrapperCodegen.generate", - "ts": 1727975456837466.0, - "args": null, + "ts": 1733527487153239.2, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.840000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "4f0fc9e676bebc5b3ec366bb1e137c82"} +V1206 15:24:47.154000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "dcff0d1fc8b779e008cb24ec2e74e7f7"} { "name": "PythonWrapperCodegen.generate", - "ts": 1727975456840311.0, + "ts": 1733527487154470.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 } +V1206 15:24:47.154000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "9be98be603274f80673ac2f6d241cec4"} + { + "name": "GraphLowering.codegen", + "ts": 1733527487154692.2, + "args": { + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_code": {"filename": "/tmp/oulgen/tmp4z1i5ywe/kp/ckpysuucou6gm55terbvpynnfevubjpwkfm3ubzxyauw2bgggpvi.py"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "187381168c614ba1181a424c67fdb77c"} +V1206 15:24:47.155000 1667746 torch/_inductor/graph.py:2030] {"inductor_output_code": {"filename": "/tmp/tmpsn77tlsm/sm/csmb7f23sz2oa7mngdcoprpgrrbziqzirwhkjffdmxv6qhdf3e57.py"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "9ced76d2b85dd84790793e85fcadcc28"} # AOT ID: ['0_inference'] from ctypes import c_void_p, c_long, c_int import torch @@ -772,7 +1065,14 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c from torch._C import _cuda_getCurrentRawStream as get_raw_stream import triton import triton.language as tl - from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph + from torch._inductor.runtime.triton_heuristics import ( + grid, + split_scan_grid, + grid_combo_kernels, + start_graph, + end_graph, + cooperative_reduction_grid, + ) aten = torch.ops.aten inductor_ops = torch.ops.inductor @@ -784,14 +1084,15 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor alloc_from_pool = torch.ops.inductor._alloc_from_pool async_compile = AsyncCompile() + empty_strided_p2p = torch._C._distributed_c10d._SymmetricMemory.empty_strided_p2p - # kernel path: /tmp/oulgen/tmp4z1i5ywe/ke/ckedjitfju7kpaxnpatsaz3gzkdz6znc4u5v2e22w2xzbjdvecy5.py + # kernel path: /tmp/tmpsn77tlsm/xo/cxokkamvqulf4elvqs25iez7wbmqqfbp3i4qy7g2hwvher2zrziz.py # Topologically Sorted Source Nodes: [flex_attention], Original ATen: [] # Source node to ATen node mapping: # flex_attention => flex_attention # Graph fragment: - # %flex_attention : [num_users=1] = call_function[target=torch.ops.higher_order.flex_attention](args = (%arg0_1, %arg1_1, %arg2_1, %sdpa_score0, (%arg3_1, %arg4_1, %arg5_1, %arg6_1, %arg7_1, %arg8_1, %arg9_1, %arg10_1, 128, 128, %sdpa_mask0), 0.125, {ROWS_GUARANTEED_SAFE: False, PRESCALE_QK: False, OUTPUT_LOGSUMEXP: False}, (), ()), kwargs = {}) + # %flex_attention : [num_users=1] = call_function[target=torch.ops.higher_order.flex_attention](args = (%arg0_1, %arg1_1, %arg2_1, %sdpa_score0, (%arg3_1, %arg4_1, %arg5_1, %arg6_1, %arg7_1, %arg8_1, %arg9_1, %arg10_1, 128, 128, %sdpa_mask0), 0.125, {PRESCALE_QK: False, ROWS_GUARANTEED_SAFE: False, BLOCKS_ARE_CONTIGUOUS: False, OUTPUT_LOGSUMEXP: True}, (), ()), kwargs = {}) triton_tem_fused_0 = async_compile.triton('triton_tem_fused_0', ''' import triton import triton.language as tl @@ -799,19 +1100,20 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c from torch._inductor.runtime import triton_helpers, triton_heuristics from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math - from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties + from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, DeviceProperties @triton_heuristics.template( num_stages=3, num_warps=4, - triton_meta={'signature': {'arg_Q': '*fp32', 'arg_K': '*fp32', 'arg_V': '*fp32', 'arg_LSE': '*fp32', 'arg_KV_NUM_BLKS': '*i32', 'arg_KV_IDX': '*i32', 'arg_FULL_KV_NUM_BLKS': '*i32', 'arg_FULL_KV_IDX': '*i32', 'out_ptr0': '*fp32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108, warp_size=32), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]}, - inductor_meta={'kernel_name': 'triton_tem_fused_0', 'backend_hash': 'FB2CA426CF35F271C56C0D69873498391AC248E25890F2B631CA8B52D56952BD', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': False, 'autotune_pointwise': True, 'autotune_remote_cache': False, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}, + triton_meta={'signature': {'arg_Q': '*fp32', 'arg_K': '*fp32', 'arg_V': '*fp32', 'arg_LSE': '*fp32', 'arg_KV_NUM_BLKS': '*i32', 'arg_KV_IDX': '*i32', 'arg_FULL_KV_NUM_BLKS': '*i32', 'arg_FULL_KV_IDX': '*i32', 'out_ptr0': '*fp32'}, 'device': DeviceProperties(type='cuda', index=0, cc=90, major=9, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=132, warp_size=32), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]}, + inductor_meta={'kernel_name': 'triton_tem_fused_0', 'backend_hash': '562E840D41CEB1D8E51DE726EA7592B0C37A0C6FBD72CF4E958863CEC11D41A7', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': False, 'autotune_pointwise': True, 'autotune_remote_cache': False, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}, ) @triton.jit def triton_tem_fused_0(arg_Q, arg_K, arg_V, arg_LSE, arg_KV_NUM_BLKS, arg_KV_IDX, arg_FULL_KV_NUM_BLKS, arg_FULL_KV_IDX, out_ptr0): - ROWS_GUARANTEED_SAFE : tl.constexpr = False PRESCALE_QK : tl.constexpr = False - OUTPUT_LOGSUMEXP : tl.constexpr = False + ROWS_GUARANTEED_SAFE : tl.constexpr = False + BLOCKS_ARE_CONTIGUOUS : tl.constexpr = False + OUTPUT_LOGSUMEXP : tl.constexpr = True FLOAT32_PRECISION : tl.constexpr = 'ieee' IS_DIVISIBLE : tl.constexpr = True SM_SCALE : tl.constexpr = 0.125 @@ -859,6 +1161,8 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c # about 20% more numerical error, but slightly faster. # ROWS_GUARANTEED_SAFE: Is it guaranteed that at least one value in each row # is not masked out? If so, we can skip an extra safety check + # BLOCKS_ARE_CONTIGUOUS: Is it guaranteed that all blocks in the mask are + # contiguous? If so, we don't need to do an indirect jump for every block tl.static_assert(SPARSE_Q_BLOCK_SIZE >= BLOCK_M and SPARSE_Q_BLOCK_SIZE % BLOCK_M == 0) tl.static_assert(SPARSE_KV_BLOCK_SIZE >= BLOCK_N and SPARSE_KV_BLOCK_SIZE % BLOCK_N == 0) @@ -1024,11 +1328,10 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c idx_d = tl.arange(0, V_HEAD_DIM)[None, :] mask = idx_m < Q_LEN - # TODO generalize and add proper mask support - xindex = idx_d + (64*idx_m) + (32768*idx_hq) + (131072*idx_zq) - tl.store(out_ptr0 + (tl.broadcast_to(idx_d + (64*idx_m) + (32768*idx_hq), acc.shape)), acc, mask) - # TODO dont want to write this if we dont require grad + xindex = idx_d + 64*idx_m + 32768*idx_hq + 131072*idx_zq + tl.store(out_ptr0 + (tl.broadcast_to(idx_d + 64*idx_m + 32768*idx_hq, acc.shape)), acc, mask) + if OUTPUT_LOGSUMEXP: off_hz = tl.program_id(1) l_ptrs = LSE + off_hz * Q_LEN + offs_m @@ -1055,9 +1358,10 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c IS_FULL_BLOCKS, ): # Redefines all kernel parameters (BLOCK_M, etc.) so we don't need to plumb them all through - ROWS_GUARANTEED_SAFE : tl.constexpr = False PRESCALE_QK : tl.constexpr = False - OUTPUT_LOGSUMEXP : tl.constexpr = False + ROWS_GUARANTEED_SAFE : tl.constexpr = False + BLOCKS_ARE_CONTIGUOUS : tl.constexpr = False + OUTPUT_LOGSUMEXP : tl.constexpr = True FLOAT32_PRECISION : tl.constexpr = 'ieee' IS_DIVISIBLE : tl.constexpr = True SM_SCALE : tl.constexpr = 0.125 @@ -1109,7 +1413,7 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c # update pointers offset = get_offset_for_next_block( start_n, kv_indices, kv_num_blocks, - SPARSE_KV_BLOCK_SIZE, SPARSE_KV_MULTIPLE, BLOCK_N + SPARSE_KV_BLOCK_SIZE, SPARSE_KV_MULTIPLE, BLOCK_N, BLOCKS_ARE_CONTIGUOUS ) V_block_ptr = tl.advance(V_block_ptr, (offset, 0)) @@ -1121,13 +1425,18 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c @triton.jit - def get_offset_for_next_block(loop_iter, col_indices, total_blocks, SPARSE_BLOCK, SPARSE_BLOCK_MULTIPLE, BLOCK): + def get_offset_for_next_block( + loop_iter, col_indices, total_blocks, + SPARSE_BLOCK, SPARSE_BLOCK_MULTIPLE, BLOCK, + BLOCKS_ARE_CONTIGUOUS: tl.constexpr + ): + if BLOCKS_ARE_CONTIGUOUS: + return BLOCK cur_block_idx = loop_iter // SPARSE_BLOCK_MULTIPLE cur_block = tl.load(col_indices + cur_block_idx, eviction_policy="evict_last") next_block = tl.load(col_indices + cur_block_idx + 1, eviction_policy="evict_last", mask=cur_block_idx + 1 < total_blocks) needs_jump = (loop_iter + 1) % SPARSE_BLOCK_MULTIPLE == 0 jump_to_block = (next_block - cur_block ) * SPARSE_BLOCK - (SPARSE_BLOCK_MULTIPLE - 1) * BLOCK - offset = jump_to_block * needs_jump + (1 - needs_jump) * BLOCK return offset @@ -1143,9 +1452,10 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c IS_FULL_BLOCKS, CHECK_BLOCK_BOUNDARY=False, ): # Redefines all kernel parameters (BLOCK_M, etc.) so we don't need to plumb them all through - ROWS_GUARANTEED_SAFE : tl.constexpr = False PRESCALE_QK : tl.constexpr = False - OUTPUT_LOGSUMEXP : tl.constexpr = False + ROWS_GUARANTEED_SAFE : tl.constexpr = False + BLOCKS_ARE_CONTIGUOUS : tl.constexpr = False + OUTPUT_LOGSUMEXP : tl.constexpr = True FLOAT32_PRECISION : tl.constexpr = 'ieee' IS_DIVISIBLE : tl.constexpr = True SM_SCALE : tl.constexpr = 0.125 @@ -1195,7 +1505,7 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c if CHECK_BLOCK_BOUNDARY: - mask_mod_output = tl.where(offs_n < KV_LEN, mask_mod_output, float("-inf")) + mask_mod_output = tl.where(offs_n < KV_LEN, mask_mod_output, False) # apply mask for partially unmasked blocks post_mod_scores = tl.where(mask_mod_output, post_mod_scores, float("-inf")) @@ -1233,7 +1543,7 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c return acc, l_i, m_i ''', device_str='cuda') - meta0 = {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False, 'FLOAT32_PRECISION': "'ieee'", 'IS_DIVISIBLE': True, 'SM_SCALE': 0.125, 'GQA_SHARED_HEADS': 1, 'HAS_FULL_BLOCKS': True, 'QK_HEAD_DIM': 64, 'V_HEAD_DIM': 64, 'BLOCK_M': 128, 'BLOCK_N': 32, 'SPARSE_Q_BLOCK_SIZE': 128, 'SPARSE_KV_BLOCK_SIZE': 128} + meta0 = {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True, 'FLOAT32_PRECISION': "'ieee'", 'IS_DIVISIBLE': True, 'SM_SCALE': 0.125, 'GQA_SHARED_HEADS': 1, 'HAS_FULL_BLOCKS': True, 'QK_HEAD_DIM': 64, 'V_HEAD_DIM': 64, 'BLOCK_M': 128, 'BLOCK_N': 32, 'SPARSE_Q_BLOCK_SIZE': 128, 'SPARSE_KV_BLOCK_SIZE': 128} async_compile.wait(globals()) @@ -1293,46 +1603,111 @@ V1003 10:10:56.841000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c from torch._inductor.wrapper_benchmark import compiled_module_main compiled_module_main('None', benchmark_compiled_module) -V1003 10:11:02.564000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "a51341d292b7eede808d8a68c1fd524e"} +V1206 15:24:47.155000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "495b4d06507b7298850f179495a17b02"} + { + "name": "PyCodeCache.load_by_key_path", + "ts": 1733527487155553.8, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:47.173000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "6dee17cfbaa0ca0649062deba0d7d459"} + { + "name": "async_compile.wait", + "ts": 1733527487173738.0, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.157000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "2d51e2d3dcba45c4d0d4d29f299f024a"} + { + "name": "async_compile.wait", + "ts": 1733527490157741.8, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.158000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0b68c334c8533667a781e2f662968ca9"} + { + "name": "PyCodeCache.load_by_key_path", + "ts": 1733527490158210.8, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.158000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "60e7a41770b95fe1d25ce4274e02d3ef"} { "name": "code_gen", - "ts": 1727975462564030.8, + "ts": 1733527490158517.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 + "fn_name": "GraphLowering.compile_to_module", + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 } +V1206 15:24:50.158000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "02264d8aae3ebc56684debf0c3f9f4de"} + { + "name": "GraphLowering.compile_to_fn", + "ts": 1733527490158861.8, + "args": { + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.564000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0723c01321e1ba5c17eafbd9000d9e2d"} +V1206 15:24:50.251000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "a469a724b0141b56eff140f8388e6f2b"} { - "name": "GraphLowering.compile_to_module", - "ts": 1727975462564643.5, + "name": "TritonBundler.collect", + "ts": 1733527490250938.0, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 } +V1206 15:24:50.252000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b19aaa17a3d9b352d63772e3deb35b7e"} + { + "name": "TritonBundler.collect", + "ts": 1733527490252606.0, + "args": { + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.603000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "489dd15ea036ef0a9a0677625f691d7c"} +V1206 15:24:50.256000 1667746 torch/_dynamo/utils.py:1327] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c0c67812e1ca34090ce7aea9849e792e"} { "name": "fx_graph_cache_miss", - "ts": 1727975456229537.0, + "ts": 1733527486241755.2, "args": { - "key": "f4lkea5y7lzhlshohvr3aqpd7bchdflfs7j5wn7mrurponawoutk", + "key": "ff6i4wr5gqsaog5zgs3qrwebcodgoghruloxh5pzs6iul3cmayau", "components": [ - "[n7x23yy6fih6vdcjzlzbhy3d6vx3ilu7ylp3zz6wkultu4yvnzn] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[v4wl2w2eawehu2yuzv4ftucrovobmkotky2oimjdvonwbz3g4ir] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", @@ -1345,6 +1720,7 @@ V1003 10:11:02.603000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", + "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", @@ -1352,7 +1728,6 @@ V1003 10:11:02.603000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", - "[f44ag5aflby2bkxl7a4k6whljrk7jat7bmreuxklei4p3czhk7p] fx_kwargs[user_visible_outputs]: {'getitem': None}", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", @@ -1366,258 +1741,270 @@ V1003 10:11:02.603000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", - "[7uhqwjfn75ek3woo3k7em2mluon5hx2ojvzlevlvjvz6xfxjhzl] torch_version: ", - "[c3z7bmoxyo6gl5hi47v6dc7jwsl55b3asd75nr25uyengi5ah3p] system_info[device]: {'name': 'NVIDIA PG509-210'}", - "[3fb7kae6ogkdd4zcm3fkjoipdpybxhn4aoxzv7z7xsfwq233e4l] system_info[version]: {'triton': '3.1.0+5fe38ffd73dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-a8fb7be728d460b7ec64ab62edb8af1bbca8994fd718cde7178d46bad64530a1-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.0'}", - "[z5x2bdhir5lzlbti73vdbfulnuu5vinzpwgmmgf4rjb775tzl3h] system_info[hash]: 9698c97edde4a99a2f3b54bbd0db5291bbcdb75c83acb376ccff61fb0bf0ac1a", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[abi_compatible]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_dump_consts_bin]: False", - "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", - "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", - "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", - "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[bundled_autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", - "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", - "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", - "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", - "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", - "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", - "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", - "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", - "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", - "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cuda.generate_test_runner]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", - "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", - "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[enable_auto_functionalized_v2]: False", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", - "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", - "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", - "[62lrdx35b7hnumwb7mp5oc5y5csm2abylvtdzfloct3noaqov3n] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'post_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", - "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", - "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", - "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", - "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", - "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", - "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", - "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", - "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", - "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", + "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", - "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", - "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_post_pass]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_pre_pass]: None", - "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", - "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", - "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", - "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", - "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", - "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_peak_memory]: False", - "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", - "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", - "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", - "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", + "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", - "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", - "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", - "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", - "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", - "[vzzema5ityqj2wepdmkulue7q5pcevdr5h27oxxutf35d4tjume] inductor_config[triton_kernel_default_layout_constraint]: flexible_layout", - "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", - "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False" + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", + "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" ], - "cache_event_time": 1727975456229537044, + "cache_event_time": 1733527486949451176, "cache_state": "miss", - "time_taken_ns": 6372999079 + "triton_bundler_meta": "TritonBundlerMetadata(cached_kernel_names=['triton_tem_fused_0'])", + "time_taken_ns": 4009141776, + "compile_id": "1/0" }, "ph": "i", "cat": "dynamo_timed", @@ -1625,140 +2012,430 @@ V1003 10:11:02.603000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "pid": 0, "s": "p" } -V1003 10:11:02.604000 2235078 torch/_inductor/codecache.py:1463] {"artifact": {"name": "fx_graph_cache_miss", "encoding": "json"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "785133baf5b9f8dad73506774a2a34dc"} - {"key": "f4lkea5y7lzhlshohvr3aqpd7bchdflfs7j5wn7mrurponawoutk", "components": ["[n7x23yy6fih6vdcjzlzbhy3d6vx3ilu7ylp3zz6wkultu4yvnzn] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[3]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[4]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[5]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[6]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[7]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[8]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", "[f44ag5aflby2bkxl7a4k6whljrk7jat7bmreuxklei4p3czhk7p] fx_kwargs[user_visible_outputs]: {'getitem': None}", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", "[kcuxe2zwm3mzv2uk6adm6iskoy35bqfv725twacrdewod2dbl5d] inputs_to_check[3]: 3", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inputs_to_check[4]: 4", "[qs5hilycp4ew4ivtc7m5jaxp7q4pm5slioxw3fi3ur6ei65ybz4] inputs_to_check[5]: 5", "[agkvbkaha53nbz3aeeuhvxjvvc4glhfjofzkg6g2qjoo2e5otcx] inputs_to_check[6]: 6", "[j3s5elu6itwgjafc7rzhy4whrbufl6kfmlufjhh25grt643bk5f] inputs_to_check[7]: 7", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inputs_to_check[8]: 8", "[qlgfiyqewrmkgqth2qm6wkq2ja5lzkapg3ypgnvoyfqqnidaoj3] inputs_to_check[9]: 9", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", "[7uhqwjfn75ek3woo3k7em2mluon5hx2ojvzlevlvjvz6xfxjhzl] torch_version: ", "[c3z7bmoxyo6gl5hi47v6dc7jwsl55b3asd75nr25uyengi5ah3p] system_info[device]: {'name': 'NVIDIA PG509-210'}", "[3fb7kae6ogkdd4zcm3fkjoipdpybxhn4aoxzv7z7xsfwq233e4l] system_info[version]: {'triton': '3.1.0+5fe38ffd73dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-a8fb7be728d460b7ec64ab62edb8af1bbca8994fd718cde7178d46bad64530a1-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.0'}", "[z5x2bdhir5lzlbti73vdbfulnuu5vinzpwgmmgf4rjb775tzl3h] system_info[hash]: 9698c97edde4a99a2f3b54bbd0db5291bbcdb75c83acb376ccff61fb0bf0ac1a", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[abi_compatible]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_dump_consts_bin]: False", "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cuda.generate_test_runner]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[enable_auto_functionalized_v2]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", "[62lrdx35b7hnumwb7mp5oc5y5csm2abylvtdzfloct3noaqov3n] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'post_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_pre_pass]: None", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_peak_memory]: False", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", "[vzzema5ityqj2wepdmkulue7q5pcevdr5h27oxxutf35d4tjume] inductor_config[triton_kernel_default_layout_constraint]: flexible_layout", "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False"], "cache_event_time": 1727975456229537044, "cache_state": "miss", "time_taken_ns": 6372999079} -V1003 10:11:02.605000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "20079c82f956785a2eeb02be390ae599"} +V1206 15:24:50.257000 1667746 torch/_inductor/compile_fx.py:751] {"artifact": {"name": "fx_graph_cache_miss", "encoding": "json"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0ad616f7fe717a86ed9e39486f7955e2"} + {"key": "ff6i4wr5gqsaog5zgs3qrwebcodgoghruloxh5pzs6iul3cmayau", "components": ["[v4wl2w2eawehu2yuzv4ftucrovobmkotky2oimjdvonwbz3g4ir] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[3]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[4]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[5]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[6]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[7]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[8]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", "[kcuxe2zwm3mzv2uk6adm6iskoy35bqfv725twacrdewod2dbl5d] inputs_to_check[3]: 3", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inputs_to_check[4]: 4", "[qs5hilycp4ew4ivtc7m5jaxp7q4pm5slioxw3fi3ur6ei65ybz4] inputs_to_check[5]: 5", "[agkvbkaha53nbz3aeeuhvxjvvc4glhfjofzkg6g2qjoo2e5otcx] inputs_to_check[6]: 6", "[j3s5elu6itwgjafc7rzhy4whrbufl6kfmlufjhh25grt643bk5f] inputs_to_check[7]: 7", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inputs_to_check[8]: 8", "[qlgfiyqewrmkgqth2qm6wkq2ja5lzkapg3ypgnvoyfqqnidaoj3] inputs_to_check[9]: 9", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[bundled_autotune_remote_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None"], "cache_event_time": 1733527486949451176, "cache_state": "miss", "triton_bundler_meta": "TritonBundlerMetadata(cached_kernel_names=['triton_tem_fused_0'])", "time_taken_ns": 4009141776, "compile_id": "1/0"} +V1206 15:24:50.257000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c7114cd61939a8286e15f75f4b1dcb23"} { "name": "inductor_compile", - "ts": 1727975462605207.0, - "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } - }, - "ph": "E", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0 - } -V1003 10:11:02.605000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "194073b2ec42e1f2dbb4aff2a005e416"} - { - "name": "compile_fx_inner", - "ts": 1727975462605545.2, + "ts": 1733527490257456.0, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "compile_fx_inner", + "compile_id": "1/0", + "is_backward": false, + "cache_state": "miss", + "cache_event_time": 1733527486241755330, + "key": "ff6i4wr5gqsaog5zgs3qrwebcodgoghruloxh5pzs6iul3cmayau", + "components": [ + "[v4wl2w2eawehu2yuzv4ftucrovobmkotky2oimjdvonwbz3g4ir] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[3]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[4]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[5]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[6]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[7]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[8]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", + "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", + "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", + "[kcuxe2zwm3mzv2uk6adm6iskoy35bqfv725twacrdewod2dbl5d] inputs_to_check[3]: 3", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inputs_to_check[4]: 4", + "[qs5hilycp4ew4ivtc7m5jaxp7q4pm5slioxw3fi3ur6ei65ybz4] inputs_to_check[5]: 5", + "[agkvbkaha53nbz3aeeuhvxjvvc4glhfjofzkg6g2qjoo2e5otcx] inputs_to_check[6]: 6", + "[j3s5elu6itwgjafc7rzhy4whrbufl6kfmlufjhh25grt643bk5f] inputs_to_check[7]: 7", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inputs_to_check[8]: 8", + "[qlgfiyqewrmkgqth2qm6wkq2ja5lzkapg3ypgnvoyfqqnidaoj3] inputs_to_check[9]: 9", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", + "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", + "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[bundled_autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", + "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", + "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", + "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", + "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", + "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", + "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", + "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" + ], + "cache_bypass_reason": null, + "remote_cache_enabled": false, + "local_cache_enabled": true }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.606000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ae70f86dc28b7efeea5cae51d87a7327"} +V1206 15:24:50.258000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "df504ff9f036fd058109eb7eba60947f"} { "name": "compile_fx..fw_compiler_base", - "ts": 1727975462606301.8, + "ts": 1733527490258109.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.609000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "7d733902132e302c6ed3910c31586fe0"} +V1206 15:24:50.259000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "8774d579b741586b86c16bfd883c37fc"} { "name": "create_aot_dispatcher_function", - "ts": 1727975462609845.0, + "ts": 1733527490259870.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.610000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c2cdcd6e0e2a687934149efb13d13ca8"} +V1206 15:24:50.260000 1667746 torch/_dynamo/utils.py:1327] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "dee620e7977ddcc771eb4099546c78e8"} { - "name": "backend_compile", - "ts": 1727975462610479.8, + "name": "autograd_cache_bypass", + "ts": 1733527486087831.0, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "cache_bypass_reason": "Unsupported call_function target flex_attention. \n Function module: torch.ops.higher_order, \nFunction name: flex_attention", + "cache_bypass_hard_exception": false, + "key": null, + "cache_state": "bypass", + "components": [], + "compile_id": "1/0" }, - "ph": "E", + "ph": "i", "cat": "dynamo_timed", "tid": 0, - "pid": 0 + "pid": 0, + "s": "p" } -V1003 10:11:02.610000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d20e4a17a90e3e33d30bd927bed546fc"} +V1206 15:24:50.260000 1667746 torch/_functorch/_aot_autograd/autograd_cache.py:763] {"artifact": {"name": "aotautograd_cache_hash", "encoding": "json"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "5601d02186053adcc1ba29fd248c1d20"} + {"cache_bypass_reason": "Unsupported call_function target flex_attention. \n Function module: torch.ops.higher_order, \nFunction name: flex_attention", "cache_bypass_hard_exception": false, "key": null, "cache_state": "bypass", "components": [], "compile_id": "1/0"} +V1206 15:24:50.260000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "1b940c0ffee7b576f294d159fb3b7a84"} { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975462610762.5, + "name": "backend_compile", + "ts": 1733527490260502.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "1/0", + "requires_subclass_dispatch": false, + "dispatch_mode": "inference", + "cache_state": "bypass", + "cache_event_time": 1733527486087830904, + "key": null, + "components": [], + "cache_bypass_reason": "Unsupported call_function target flex_attention. \n Function module: torch.ops.higher_order, \nFunction name: flex_attention", + "remote_cache_enabled": false, + "local_cache_enabled": true }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.642000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_str": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "35a49d1f2da1768efdba707d805b4b97"} +V1206 15:24:50.275000 1667746 torch/_dynamo/guards.py:2315] {"dynamo_cpp_guards_str": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "004a7ba640839d40992c1546206d56ff"} TREE_GUARD_MANAGER: +- RootGuardManager - | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:471 in init_ambient_guards + | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:484 in init_ambient_guards | +- GLOBAL_STATE: ___check_global_state() | +- TORCH_FUNCTION_MODE_STACK: ___check_torch_function_mode_stack() - | +- GuardManager: source=L['k'], accessed_by=DictGetItemGuardAccessor(k) - | | +- TYPE_MATCH: ___check_type_id(L['k'], 82028112) + | +- GuardManager: source=L['k'], accessed_by=DictGetItemGuardAccessor('k') + | | +- TYPE_MATCH: ___check_type_id(L['k'], 82181376) | | +- TENSOR_MATCH: check_tensor(L['k'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) | | +- NO_HASATTR: hasattr(L['k'], '_dynamo_dynamic_indices') == False | | +- NO_TENSOR_ALIASING: check_no_aliasing(L['k'], L['q'], L['v'], L['block_mask'].q_indices, L['block_mask'].kv_indices, L['block_mask'].q_num_blocks, L['block_mask'].kv_num_blocks, L['block_mask'].full_q_indices, L['block_mask'].full_kv_indices, L['block_mask'].full_q_num_blocks, L['block_mask'].full_kv_num_blocks) - | +- GuardManager: source=L['q'], accessed_by=DictGetItemGuardAccessor(q) - | | +- TYPE_MATCH: ___check_type_id(L['q'], 82028112) + | | +- GuardManager: source=L['k'].dim, accessed_by=GetAttrGuardAccessor(dim) + | | +- GuardManager: source=L['k'].size, accessed_by=GetAttrGuardAccessor(size) + | +- GuardManager: source=L['q'], accessed_by=DictGetItemGuardAccessor('q') + | | +- TYPE_MATCH: ___check_type_id(L['q'], 82181376) | | +- TENSOR_MATCH: check_tensor(L['q'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) | | +- NO_HASATTR: hasattr(L['q'], '_dynamo_dynamic_indices') == False | | +- NO_TENSOR_ALIASING - | +- GuardManager: source=L['v'], accessed_by=DictGetItemGuardAccessor(v) - | | +- TYPE_MATCH: ___check_type_id(L['v'], 82028112) + | | +- GuardManager: source=L['q'].dim, accessed_by=GetAttrGuardAccessor(dim) + | | +- GuardManager: source=L['q'].size, accessed_by=GetAttrGuardAccessor(size) + | +- GuardManager: source=L['v'], accessed_by=DictGetItemGuardAccessor('v') + | | +- TYPE_MATCH: ___check_type_id(L['v'], 82181376) | | +- TENSOR_MATCH: check_tensor(L['v'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) | | +- NO_HASATTR: hasattr(L['v'], '_dynamo_dynamic_indices') == False | | +- NO_TENSOR_ALIASING - | +- GuardManager: source=L['score_mod'], accessed_by=DictGetItemGuardAccessor(score_mod) + | | +- GuardManager: source=L['v'].dim, accessed_by=GetAttrGuardAccessor(dim) + | | +- GuardManager: source=L['v'].size, accessed_by=GetAttrGuardAccessor(size) + | +- GuardManager: source=L['score_mod'], accessed_by=DictGetItemGuardAccessor('score_mod') | | +- GuardManager: source=L['score_mod'].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | +- ID_MATCH: ___check_obj_id(L['score_mod'].__code__, 140413271879296) - | +- GuardManager: source=L['block_mask'], accessed_by=DictGetItemGuardAccessor(block_mask) - | | +- TYPE_MATCH: ___check_type_id(L['block_mask'], 387600320) + | | | +- ID_MATCH: ___check_obj_id(L['score_mod'].__code__, 140062268556144) + | +- GuardManager: source=L['block_mask'], accessed_by=DictGetItemGuardAccessor('block_mask') + | | +- TYPE_MATCH: ___check_type_id(L['block_mask'], 139679664) | | +- GuardManager: source=L['block_mask'].mask_mod, accessed_by=GetAttrGuardAccessor(mask_mod) | | | +- GuardManager: source=L['block_mask'].mask_mod.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].mask_mod.__code__, 140413271880128) + | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].mask_mod.__code__, 140062268556768) | | +- GuardManager: source=L['block_mask'].q_indices, accessed_by=GetAttrGuardAccessor(q_indices) | | | +- TENSOR_MATCH: check_tensor(L['block_mask'].q_indices, Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.int32, device=0, requires_grad=False, size=[1, 1, 16, 16], stride=[256, 256, 16, 1]) | | | +- NO_HASATTR: hasattr(L['block_mask'].q_indices, '_dynamo_dynamic_indices') == False | | | +- NO_TENSOR_ALIASING | | +- GuardManager: source=L['block_mask'].BLOCK_SIZE, accessed_by=GetAttrGuardAccessor(BLOCK_SIZE) - | | | +- TYPE_MATCH: ___check_type_id(L['block_mask'].BLOCK_SIZE, 8815232) + | | | +- TYPE_MATCH: ___check_type_id(L['block_mask'].BLOCK_SIZE, 8812224) | | | +- LENGTH_CHECK: len(L['block_mask'].BLOCK_SIZE) == 2 | | | +- GuardManager: source=L['block_mask'].BLOCK_SIZE[0], accessed_by=TupleGetItemGuardAccessor(0) | | | | +- EQUALS_MATCH: L['block_mask'].BLOCK_SIZE[0] == 128 @@ -1795,48 +2472,54 @@ V1003 10:11:02.642000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | +- GuardManager: source=L['block_mask'].as_tuple, accessed_by=GetAttrGuardAccessor(as_tuple) | | | +- GuardManager: source=L['block_mask'].as_tuple, accessed_by=FuncDefaultsGuardAccessor | | | | +- GuardManager: source=L['block_mask'].as_tuple.__defaults__[0], accessed_by=GetItemGuardAccessor(0) - | | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].as_tuple.__defaults__[0], 8911040) - | +- GuardManager: source=L['flex_attention'], accessed_by=DictGetItemGuardAccessor(flex_attention) + | | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].as_tuple.__defaults__[0], 8908032) + | +- GuardManager: source=L['flex_attention'], accessed_by=DictGetItemGuardAccessor('flex_attention') | | +- GuardManager: source=L['flex_attention'].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__code__, 387082992) + | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__code__, 139318784) | | +- GuardManager: source=L['flex_attention'], accessed_by=FuncDefaultsGuardAccessor | | | +- GuardManager: source=L['flex_attention'].__defaults__[2], accessed_by=GetItemGuardAccessor(2) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[2], 8825760) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[2], 8822752) | | | +- GuardManager: source=L['flex_attention'].__defaults__[3], accessed_by=GetItemGuardAccessor(3) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[3], 8910592) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[3], 8907584) | | | +- GuardManager: source=L['flex_attention'].__defaults__[4], accessed_by=GetItemGuardAccessor(4) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[4], 8910592) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[4], 8907584) | | | +- GuardManager: source=L['flex_attention'].__defaults__[5], accessed_by=GetItemGuardAccessor(5) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[5], 8825760) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[5], 8822752) | +- GuardManager: source=G, accessed_by=GlobalsGuardAccessor - | | +- GuardManager: source=G['__builtins_dict___2'], accessed_by=DictGetItemGuardAccessor(__builtins_dict___2) - | | | +- GuardManager: source=G['__builtins_dict___2']['len'], accessed_by=DictGetItemGuardAccessor(len) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['len'], 140413275558816) - | | | +- GuardManager: source=G['__builtins_dict___2']['sum'], accessed_by=DictGetItemGuardAccessor(sum) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['sum'], 140413275559936) - | | | +- GuardManager: source=G['__builtins_dict___2']['list'], accessed_by=DictGetItemGuardAccessor(list) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['list'], 8844320) - | | | +- GuardManager: source=G['__builtins_dict___2']['type'], accessed_by=DictGetItemGuardAccessor(type) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['type'], 8813248) - | | | +- GuardManager: source=G['__builtins_dict___2']['tuple'], accessed_by=DictGetItemGuardAccessor(tuple) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['tuple'], 8815232) - | | | +- GuardManager: source=G['__builtins_dict___2']['object'], accessed_by=DictGetItemGuardAccessor(object) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['object'], 8813984) - | | | +- GuardManager: source=G['__builtins_dict___2']['isinstance'], accessed_by=DictGetItemGuardAccessor(isinstance) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['isinstance'], 140413275558496) - | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_utils_dot__pytree) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140411217627952) + | | +- GuardManager: source=G['_140062263790704_c1'], accessed_by=DictGetItemGuardAccessor('_140062263790704_c1') + | | | +- GuardManager: source=G['_140062263790704_c1'].Tensor, accessed_by=GetAttrGuardAccessor(Tensor) + | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c1'].Tensor, 82181376) + | | | | +- GuardManager: source=G['_140062263790704_c1'].Tensor.__bases__, accessed_by=GetAttrGuardAccessor(__bases__) + | | | | | +- GuardManager: source=G['_140062263790704_c1'].Tensor.__bases__[0], accessed_by=TupleGetItemGuardAccessor(0) + | | | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c1'].Tensor.__bases__[0], 140062119703136) + | | +- GuardManager: source=G['__builtins_dict___2'], accessed_by=DictGetItemGuardAccessor('__builtins_dict___2') + | | | +- GuardManager: source=G['__builtins_dict___2']['len'], accessed_by=DictGetItemGuardAccessor('len') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['len'], 140062269592480) + | | | +- GuardManager: source=G['__builtins_dict___2']['sum'], accessed_by=DictGetItemGuardAccessor('sum') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['sum'], 140062269593600) + | | | +- GuardManager: source=G['__builtins_dict___2']['list'], accessed_by=DictGetItemGuardAccessor('list') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['list'], 8841312) + | | | +- GuardManager: source=G['__builtins_dict___2']['type'], accessed_by=DictGetItemGuardAccessor('type') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['type'], 8810240) + | | | +- GuardManager: source=G['__builtins_dict___2']['tuple'], accessed_by=DictGetItemGuardAccessor('tuple') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['tuple'], 8812224) + | | | +- GuardManager: source=G['__builtins_dict___2']['object'], accessed_by=DictGetItemGuardAccessor('object') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['object'], 8810976) + | | | +- GuardManager: source=G['__builtins_dict___2']['isinstance'], accessed_by=DictGetItemGuardAccessor('isinstance') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___2']['isinstance'], 140062269592160) + | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_utils_dot__pytree') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140057600329744) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].TreeSpec, accessed_by=GetAttrGuardAccessor(TreeSpec) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 84866496) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 86791872) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf, accessed_by=GetAttrGuardAccessor(_is_leaf) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140411217262720) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140057601749808) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, accessed_by=GetAttrGuardAccessor(_LEAF_SPEC) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 85171104) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 87025712) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, accessed_by=GetAttrGuardAccessor(type) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, accessed_by=GetAttrGuardAccessor(context) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes, accessed_by=GetAttrGuardAccessor(num_nodes) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes == 1 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_leaves, accessed_by=GetAttrGuardAccessor(num_leaves) @@ -1844,37 +2527,37 @@ V1003 10:11:02.642000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children, accessed_by=GetAttrGuardAccessor(num_children) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children == 0 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, accessed_by=GetAttrGuardAccessor(children_specs) - | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8844320) + | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8841312) | | | | | +- LENGTH_CHECK: not G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type, accessed_by=GetAttrGuardAccessor(_get_node_type) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140411217262448) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140057601749536) | | | +- DictGuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES, accessed_by=GetAttrGuardAccessor(SUPPORTED_NODES) - | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 519596 + | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 366913 | | | | +- KeyValueManager pair at index=1 | | | | | +- ValueManager: GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]] | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn, accessed_by=GetAttrGuardAccessor(flatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140411196281984) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140057599989696) | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn, accessed_by=GetAttrGuardAccessor(unflatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140411217182288) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140057600267792) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper, accessed_by=GetAttrGuardAccessor(_tree_flatten_helper) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140411217413040) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140057600001392) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance, accessed_by=GetAttrGuardAccessor(_is_namedtuple_instance) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140411217412592) - | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_comptime'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot__dynamo_dot_comptime) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_comptime'], 140410226912176) - | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_decorators'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot__dynamo_dot_decorators) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'], 140410226910096) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140057600000944) + | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_comptime'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot__dynamo_dot_comptime') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_comptime'], 140057359527872) + | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_decorators'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot__dynamo_dot_decorators') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'], 140057359526192) | | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_decorators'].is_compiling, accessed_by=GetAttrGuardAccessor(is_compiling) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'].is_compiling, 140410376252096) - | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_nn_dot_attention_dot__utils) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot__utils'], 140409673896784) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'].is_compiling, 140057422814624) + | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_nn_dot_attention_dot__utils') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot__utils'], 140048551571072) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS, accessed_by=GetAttrGuardAccessor(_SUPPORTED_HEAD_DIMS) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS, 8844320) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS, 8841312) | | | | +- LENGTH_CHECK: len(G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS) == 10 | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[0], accessed_by=ListGetItemGuardAccessor(0) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[0] == 2 @@ -1888,131 +2571,155 @@ V1003 10:11:02.642000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[4] == 32 | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[5], accessed_by=ListGetItemGuardAccessor(5) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[5] == 64 - | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_nn_dot_attention_dot_flex_attention) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], 140409673895824) + | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_nn_dot_attention_dot_flex_attention') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], 140048551568912) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math, accessed_by=GetAttrGuardAccessor(math) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math, 140413266939392) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math, 140062267404384) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math.sqrt, accessed_by=GetAttrGuardAccessor(sqrt) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math.sqrt, 140413266943072) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math.sqrt, 140062267408144) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch, accessed_by=GetAttrGuardAccessor(torch) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch, 140413267918368) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch, 140062263790704) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo, accessed_by=GetAttrGuardAccessor(_dynamo) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo, 140413260098400) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo, 140062262057760) | | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static, accessed_by=GetAttrGuardAccessor(mark_static) | | | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static.__code__, 123166432) + | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static.__code__, 125680000) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler, accessed_by=GetAttrGuardAccessor(compiler) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler, 140410826010400) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler, 140057571877776) | | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler.is_dynamo_compiling, accessed_by=GetAttrGuardAccessor(is_dynamo_compiling) - | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler.is_dynamo_compiling, 140410826132992) + | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler.is_dynamo_compiling, 140057572232544) + | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.is_grad_enabled, accessed_by=GetAttrGuardAccessor(is_grad_enabled) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.is_grad_enabled, 140062253115936) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device, accessed_by=GetAttrGuardAccessor(_validate_device) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device.__code__, 140409673611088) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device.__code__, 140048552018544) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop, accessed_by=GetAttrGuardAccessor(flex_attention_hop) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop, 96992544) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop, 96230624) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop.__name__, accessed_by=GetAttrGuardAccessor(__name__) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop.__name__ == 'flex_attention' | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim, accessed_by=GetAttrGuardAccessor(_supported_head_dim) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim.__code__, 140409673231376) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim.__code__, 140048552864992) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim, accessed_by=GetAttrGuardAccessor(_validate_embed_dim) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim.__code__, 388086512) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim.__code__, 139781872) + | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_nestedness, accessed_by=GetAttrGuardAccessor(_validate_nestedness) + | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_nestedness.__code__, accessed_by=GetAttrGuardAccessor(__code__) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_nestedness.__code__, 140048553100560) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input, accessed_by=GetAttrGuardAccessor(_validate_sdpa_input) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input.__code__, 387915104) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input.__code__, 139823744) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options, accessed_by=GetAttrGuardAccessor(_apply_kernel_options) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options.__code__, 140409683680752) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options.__code__, 140048554555888) -V1003 10:11:02.643000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "3937c61e6c98ab6b52d8e00fc2327f9d"} +V1206 15:24:50.275000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f0dcf66ffff430aab1506701a2358ee4"} { "name": "entire_frame_compile", - "ts": 1727975462642948.2, + "ts": 1733527490275762.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "_compile.compile_inner", + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.643000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "af7756663bbea703738c5d774f3ca273"} +V1206 15:24:50.278000 1667746 torch/_dynamo/utils.py:1045] {"compilation_metrics": {"pre_grad_pass_time_us": 2504, "joint_graph_pass_time_us": 100514, "feature_usage": {"pytorch/remote_cache:fx_graph_memcache_version": true, "pytorch/remote_cache:bundle_triton_into_fx_graph_cache_v2": true}, "post_grad_pass_time_us": 9066, "inductor_code_gen_cumulative_compile_time_us": 3019492, "inductor_cumulative_compile_time_us": 4016483, "aot_autograd_cumulative_compile_time_us": 4177883, "dynamo_cumulative_compile_time_us": 4298670, "frame_key": "2", "co_name": "fn", "co_filename": "/data/users/xmfan/a/pytorch/test/inductor/test_codecache.py", "co_firstlineno": 459, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 91, "shape_env_guard_count": 0, "graph_op_count": 2, "graph_node_count": 16, "graph_input_count": 11, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": false, "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}", "is_forward": true, "dynamo_compile_time_before_restart_us": 0}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.278000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "e38d71b3f183088f1bc592d1f308f959"} { - "name": "_compile.compile_inner", - "ts": 1727975462643261.5, + "name": "dynamo", + "ts": 1733527490278673.0, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0", + "frame_key": "2", + "co_name": "fn", + "co_filename": "/data/users/xmfan/a/pytorch/test/inductor/test_codecache.py", + "co_firstlineno": 459, + "cache_size": 0, + "accumulated_cache_size": 0, + "guard_count": 91, + "shape_env_guard_count": 0, + "graph_op_count": 2, + "graph_node_count": 16, + "graph_input_count": 11, + "fail_type": null, + "fail_reason": null, + "fail_user_frame_filename": null, + "fail_user_frame_lineno": null, + "non_compliant_ops": [], + "compliant_custom_ops": [], + "restart_reasons": [], + "dynamo_time_before_restart_s": 0.0, + "has_guarded_code": true, + "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.643000 2235078 torch/_dynamo/utils.py:840] {"compilation_metrics": {"compile_id": "1/0", "frame_key": "2", "co_name": "fn", "co_filename": "/data/users/oulgen/pytorch/test/inductor/test_codecache.py", "co_firstlineno": 379, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 80, "shape_env_guard_count": 0, "graph_op_count": 11, "graph_node_count": 25, "graph_input_count": 11, "start_time": 1727975453.4995801, "entire_frame_compile_time_s": 9.143243551254272, "backend_compile_time_s": 8.912311315536499, "inductor_compile_time_s": 8.53272271156311, "code_gen_time_s": 5.749065160751343, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "possibly_missed_reinplacing_opportunities": 0, "remote_cache_time_saved_s": 0, "structured_logging_overhead_s": 0.034261202, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": true}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.649000 2235078 torch/_dynamo/convert_frame.py:915] {"dynamo_start": {"stack": [{"line": 916, "name": "", "filename": 1}, {"line": 14, "name": "run_tests", "filename": 2}, {"line": 38, "name": "run_tests", "filename": 3}, {"line": 1273, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3116, "name": "run", "filename": 4}, {"line": 3088, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 2983, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 405, "name": "test_flex_attention_caching", "filename": 1}, {"line": 380, "name": "fn", "filename": 1}, {"line": 1062, "name": "flex_attention", "filename": 10}, {"line": 1049, "name": "_flex_attention_hop_wrapper", "filename": 10}]}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.649000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "a2ef97fb78450a26e1e87b56945ab331"} +V1206 15:24:50.282000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "394328a4ef4c4a7736913bac1b8ace11"} { - "name": "_compile.compile_inner", - "ts": 1727975462649375.8, - "args": null, + "name": "dynamo", + "ts": 1733527490282024.0, + "args": { + "compile_id": "0/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.649000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "3a33794e999255e5aaa28e163fcdc260"} +V1206 15:24:50.282000 1667746 torch/_dynamo/convert_frame.py:961] {"dynamo_start": {"stack": [{"line": 1439, "name": "", "filename": 1}, {"line": 15, "name": "run_tests", "filename": 2}, {"line": 39, "name": "run_tests", "filename": 3}, {"line": 1353, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3234, "name": "run", "filename": 4}, {"line": 3206, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 3099, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 483, "name": "test_flex_attention_caching", "filename": 1}, {"line": 460, "name": "fn", "filename": 1}, {"line": 1316, "name": "flex_attention", "filename": 10}, {"line": 1303, "name": "_flex_attention_hop_wrapper", "filename": 10}]}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.282000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "db937f3641ee59365a7de351869e46cf"} { "name": "entire_frame_compile", - "ts": 1727975462649375.8, - "args": null, + "ts": 1733527490282590.2, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.652000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 240, "size": 524288}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.652000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.653000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 0, "source": "L['args'][0]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.768000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 1, "describer_id": 240, "size": 524288}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.769000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 4, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.769000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 4, "source": "L['args'][1]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.770000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 2, "describer_id": 240, "size": 524288}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.770000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 5, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.771000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 5, "source": "L['args'][2]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.772000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 3, "describer_id": 240, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.772000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 6, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.772000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 6, "source": "L['args'][4][0]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.773000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 4, "describer_id": 240, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.774000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.774000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 7, "source": "L['args'][4][1]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.775000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 5, "describer_id": 240, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.775000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.775000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 8, "source": "L['args'][4][2]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.776000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 6, "describer_id": 240, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.777000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.777000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 9, "source": "L['args'][4][3]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.778000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 7, "describer_id": 240, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.778000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.778000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 10, "source": "L['args'][4][4]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.779000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 8, "describer_id": 240, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.780000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.780000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 11, "source": "L['args'][4][5]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.781000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 9, "describer_id": 240, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.781000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 12, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.781000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 12, "source": "L['args'][4][6]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.782000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 10, "describer_id": 240, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.783000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 13, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 240}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.783000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 240, "id": 13, "source": "L['args'][4][7]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.791000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_output_graph": {"sizes": {"l_args_0_": [1, 4, 512, 64], "l_args_1_": [1, 4, 512, 64], "l_args_2_": [1, 4, 512, 64], "l_args_4_0_": [1, 1, 16], "l_args_4_1_": [1, 1, 16, 16], "l_args_4_2_": [1, 1, 16], "l_args_4_3_": [1, 1, 16, 16], "l_args_4_4_": [1, 1, 16], "l_args_4_5_": [1, 1, 16, 16], "l_args_4_6_": [1, 1, 16], "l_args_4_7_": [1, 1, 16, 16], "child_1": [], "child_2": [], "child_3": [], "child_4": [], "child": [], "child_5": [], "child_6": [], "child_7": [], "child_8": [], "getitem": [1, 4, 512, 64], "getitem_1": [1, 4, 512]}}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "36de5ad6eb1efc648a27dc62c107a2ca"} +V1206 15:24:50.284000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 0, "describer_id": 228, "size": 524288}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.284000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.284000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 0, "source": "L['args'][0]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.346000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 1, "describer_id": 228, "size": 524288}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.346000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 4, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.347000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 4, "source": "L['args'][1]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.347000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 2, "describer_id": 228, "size": 524288}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.347000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 5, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.347000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 5, "source": "L['args'][2]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.348000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 3, "describer_id": 228, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.348000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 6, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.348000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 6, "source": "L['args'][4][0]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.349000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 4, "describer_id": 228, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.349000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.349000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 7, "source": "L['args'][4][1]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.350000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 5, "describer_id": 228, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.350000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.350000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 8, "source": "L['args'][4][2]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.350000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 6, "describer_id": 228, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.350000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.351000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 9, "source": "L['args'][4][3]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.351000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 7, "describer_id": 228, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.351000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.351000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 10, "source": "L['args'][4][4]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.352000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 8, "describer_id": 228, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.352000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.352000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 11, "source": "L['args'][4][5]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.353000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 9, "describer_id": 228, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.353000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 12, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.353000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 12, "source": "L['args'][4][6]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.353000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 10, "describer_id": 228, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.354000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 13, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 228}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.354000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 228, "id": 13, "source": "L['args'][4][7]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.358000 1667746 torch/_dynamo/output_graph.py:1336] {"dynamo_output_graph": {"sizes": {"l_args_0_": [1, 4, 512, 64], "l_args_1_": [1, 4, 512, 64], "l_args_2_": [1, 4, 512, 64], "l_args_4_0_": [1, 1, 16], "l_args_4_1_": [1, 1, 16, 16], "l_args_4_2_": [1, 1, 16], "l_args_4_3_": [1, 1, 16, 16], "l_args_4_4_": [1, 1, 16], "l_args_4_5_": [1, 1, 16, 16], "l_args_4_6_": [1, 1, 16], "l_args_4_7_": [1, 1, 16, 16], "getitem": [1, 4, 512, 64], "getitem_1": [1, 4, 512]}}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "52fbb526c9a1f1575361cb8ed8e7794e"} class GraphModule(torch.nn.Module): def forward(self, L_args_0_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_args_1_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_args_2_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_args_4_0_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_1_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_args_4_2_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_3_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_args_4_4_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_5_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_args_4_6_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_7_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): l_args_0_ = L_args_0_ @@ -2027,97 +2734,62 @@ V1003 10:11:02.791000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_outpu l_args_4_6_ = L_args_4_6_ l_args_4_7_ = L_args_4_7_ - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1050 in _flex_attention_hop_wrapper, code: return flex_attention_hop(*args, **kwargs) - child_1: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_1 = None - child_2: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_2 = None - child_3: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_3 = None - child_4: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_4 = None - child: "f32[][]cuda:0" = l_args_0_.new_empty([], requires_grad = False); child = None + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1304 in _flex_attention_hop_wrapper, code: return flex_attention_hop(*args, **kwargs) score_mod_0 = self.score_mod_0 - child_5: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_5 = None - child_6: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_6 = None - child_7: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_7 = None - child_8: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_8 = None mask_fn_0 = self.mask_fn_0 - flex_attention = torch.ops.higher_order.flex_attention(l_args_0_, l_args_1_, l_args_2_, score_mod_0, (l_args_4_0_, l_args_4_1_, l_args_4_2_, l_args_4_3_, l_args_4_4_, l_args_4_5_, l_args_4_6_, l_args_4_7_, 128, 128, mask_fn_0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); l_args_0_ = l_args_1_ = l_args_2_ = score_mod_0 = l_args_4_0_ = l_args_4_1_ = l_args_4_2_ = l_args_4_3_ = l_args_4_4_ = l_args_4_5_ = l_args_4_6_ = l_args_4_7_ = mask_fn_0 = None + flex_attention = torch.ops.higher_order.flex_attention(l_args_0_, l_args_1_, l_args_2_, score_mod_0, (l_args_4_0_, l_args_4_1_, l_args_4_2_, l_args_4_3_, l_args_4_4_, l_args_4_5_, l_args_4_6_, l_args_4_7_, 128, 128, mask_fn_0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); l_args_0_ = l_args_1_ = l_args_2_ = score_mod_0 = l_args_4_0_ = l_args_4_1_ = l_args_4_2_ = l_args_4_3_ = l_args_4_4_ = l_args_4_5_ = l_args_4_6_ = l_args_4_7_ = mask_fn_0 = None getitem: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0] getitem_1: "f32[1, 4, 512][2048, 512, 1]cuda:0" = flex_attention[1]; flex_attention = None return (getitem, getitem_1) class score_mod_0(torch.nn.Module): def forward(self, child: "f32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0", child_4: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:377 in score_mod, code: return score + (q - kv) + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:457 in score_mod, code: return score + (q - kv) sub: "i32[][]cuda:0" = child_3 - child_4; child_3 = child_4 = None add: "f32[][]cuda:0" = child + sub; child = sub = None return add class mask_fn_0(torch.nn.Module): - def forward(self, child_5: "i32[][]cuda:0", child_6: "i32[][]cuda:0", child_7: "i32[][]cuda:0", child_8: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 - ge: "b8[][]cuda:0" = child_7 >= child_8; child_7 = child_8 = None + def forward(self, child: "i32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0"): + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + ge: "b8[][]cuda:0" = child_2 >= child_3; child_2 = child_3 = None return ge -V1003 10:11:02.792000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "bbe4f56afd5b46c1bcfb8b89b3682d1c"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975462792636.8, - "args": null, - "ph": "B", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0 - } -V1003 10:11:02.793000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "42f674bdb6c8ba79dc796eb9e177d39e"} +V1206 15:24:50.358000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "7ea1fa27ae5b4f6ebbb5c849ee69a9a2"} { "name": "backend_compile", - "ts": 1727975462792636.8, - "args": null, - "ph": "B", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0 - } -V1003 10:11:02.793000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "70d8c05c0dc42c6de392be3605cefc5d"} - { - "name": "backend_compile", - "ts": 1727975462793358.8, + "ts": 1733527490358949.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" }, - "ph": "E", + "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.793000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "51a3e0876977448806622fe8baf64561"} +V1206 15:24:50.359000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "7f0ded16da08aebef0f53116ced241f9"} { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975462793614.8, + "name": "backend_compile", + "ts": 1733527490359226.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.818000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_str": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2b6b7ba9e9bb8df50f6d571c50ac67a8"} +V1206 15:24:50.370000 1667746 torch/_dynamo/guards.py:2315] {"dynamo_cpp_guards_str": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "217903d8e83efab45bc8d1821a378153"} TREE_GUARD_MANAGER: +- RootGuardManager - | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:471 in init_ambient_guards + | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:484 in init_ambient_guards | +- GLOBAL_STATE: ___check_global_state() | +- TORCH_FUNCTION_MODE_STACK: ___check_torch_function_mode_stack() - | +- GuardManager: source=L['args'], accessed_by=DictGetItemGuardAccessor(args) - | | +- TYPE_MATCH: ___check_type_id(L['args'], 8815232) + | +- GuardManager: source=L['args'], accessed_by=DictGetItemGuardAccessor('args') + | | +- TYPE_MATCH: ___check_type_id(L['args'], 8812224) | | +- LENGTH_CHECK: len(L['args']) == 7 | | +- GuardManager: source=L['args'][0], accessed_by=TupleGetItemGuardAccessor(0) | | | +- TENSOR_MATCH: check_tensor(L['args'][0], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) @@ -2133,9 +2805,9 @@ V1003 10:11:02.818000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | +- NO_TENSOR_ALIASING | | +- GuardManager: source=L['args'][3], accessed_by=TupleGetItemGuardAccessor(3) | | | +- GuardManager: source=L['args'][3].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][3].__code__, 140413271879296) + | | | | +- ID_MATCH: ___check_obj_id(L['args'][3].__code__, 140062268556144) | | +- GuardManager: source=L['args'][4], accessed_by=TupleGetItemGuardAccessor(4) - | | | +- TYPE_MATCH: ___check_type_id(L['args'][4], 8815232) + | | | +- TYPE_MATCH: ___check_type_id(L['args'][4], 8812224) | | | +- LENGTH_CHECK: len(L['args'][4]) == 11 | | | +- GuardManager: source=L['args'][4][0], accessed_by=TupleGetItemGuardAccessor(0) | | | | +- TENSOR_MATCH: check_tensor(L['args'][4][0], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.int32, device=0, requires_grad=False, size=[1, 1, 16], stride=[16, 16, 1]) @@ -2175,52 +2847,60 @@ V1003 10:11:02.818000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | +- EQUALS_MATCH: L['args'][4][9] == 128 | | | +- GuardManager: source=L['args'][4][10], accessed_by=TupleGetItemGuardAccessor(10) | | | | +- GuardManager: source=L['args'][4][10].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(L['args'][4][10].__code__, 140413271880128) + | | | | | +- ID_MATCH: ___check_obj_id(L['args'][4][10].__code__, 140062268556768) | | +- GuardManager: source=L['args'][5], accessed_by=TupleGetItemGuardAccessor(5) | | | +- EQUALS_MATCH: L['args'][5] == 0.125 | | +- GuardManager: source=L['args'][6], accessed_by=TupleGetItemGuardAccessor(6) - | | | +- DICT_LENGTH: len(L['args'][6]) == 3 - | | | +- GuardManager: source=L['args'][6]['ROWS_GUARANTEED_SAFE'], accessed_by=DictGetItemGuardAccessor(ROWS_GUARANTEED_SAFE) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['ROWS_GUARANTEED_SAFE'], 8910592) - | | | +- GuardManager: source=L['args'][6]['PRESCALE_QK'], accessed_by=DictGetItemGuardAccessor(PRESCALE_QK) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['PRESCALE_QK'], 8910592) - | | | +- GuardManager: source=L['args'][6]['OUTPUT_LOGSUMEXP'], accessed_by=DictGetItemGuardAccessor(OUTPUT_LOGSUMEXP) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['OUTPUT_LOGSUMEXP'], 8910592) - | +- GuardManager: source=L['kwargs'], accessed_by=DictGetItemGuardAccessor(kwargs) + | | | +- DICT_LENGTH: len(L['args'][6]) == 4 + | | | +- GuardManager: source=L['args'][6]['PRESCALE_QK'], accessed_by=DictGetItemGuardAccessor('PRESCALE_QK') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['PRESCALE_QK'], 8907584) + | | | +- GuardManager: source=L['args'][6]['ROWS_GUARANTEED_SAFE'], accessed_by=DictGetItemGuardAccessor('ROWS_GUARANTEED_SAFE') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['ROWS_GUARANTEED_SAFE'], 8907584) + | | | +- GuardManager: source=L['args'][6]['BLOCKS_ARE_CONTIGUOUS'], accessed_by=DictGetItemGuardAccessor('BLOCKS_ARE_CONTIGUOUS') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['BLOCKS_ARE_CONTIGUOUS'], 8907584) + | | | +- GuardManager: source=L['args'][6]['OUTPUT_LOGSUMEXP'], accessed_by=DictGetItemGuardAccessor('OUTPUT_LOGSUMEXP') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['OUTPUT_LOGSUMEXP'], 8908032) + | +- GuardManager: source=L['kwargs'], accessed_by=DictGetItemGuardAccessor('kwargs') | | +- DICT_LENGTH: not L['kwargs'] | +- GuardManager: source=G, accessed_by=GlobalsGuardAccessor - | | +- GuardManager: source=G['flex_attention_hop'], accessed_by=DictGetItemGuardAccessor(flex_attention_hop) - | | | +- TYPE_MATCH: ___check_type_id(G['flex_attention_hop'], 96992544) + | | +- GuardManager: source=G['flex_attention_hop'], accessed_by=DictGetItemGuardAccessor('flex_attention_hop') + | | | +- TYPE_MATCH: ___check_type_id(G['flex_attention_hop'], 96230624) | | | +- GuardManager: source=G['flex_attention_hop'].__name__, accessed_by=GetAttrGuardAccessor(__name__) | | | | +- EQUALS_MATCH: G['flex_attention_hop'].__name__ == 'flex_attention' - | | +- GuardManager: source=G['__builtins_dict___4'], accessed_by=DictGetItemGuardAccessor(__builtins_dict___4) - | | | +- GuardManager: source=G['__builtins_dict___4']['len'], accessed_by=DictGetItemGuardAccessor(len) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['len'], 140413275558816) - | | | +- GuardManager: source=G['__builtins_dict___4']['sum'], accessed_by=DictGetItemGuardAccessor(sum) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['sum'], 140413275559936) - | | | +- GuardManager: source=G['__builtins_dict___4']['list'], accessed_by=DictGetItemGuardAccessor(list) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['list'], 8844320) - | | | +- GuardManager: source=G['__builtins_dict___4']['type'], accessed_by=DictGetItemGuardAccessor(type) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['type'], 8813248) - | | | +- GuardManager: source=G['__builtins_dict___4']['tuple'], accessed_by=DictGetItemGuardAccessor(tuple) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['tuple'], 8815232) - | | | +- GuardManager: source=G['__builtins_dict___4']['object'], accessed_by=DictGetItemGuardAccessor(object) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['object'], 8813984) - | | | +- GuardManager: source=G['__builtins_dict___4']['isinstance'], accessed_by=DictGetItemGuardAccessor(isinstance) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['isinstance'], 140413275558496) - | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_utils_dot__pytree) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140411217627952) + | | +- GuardManager: source=G['_140062263790704_c2'], accessed_by=DictGetItemGuardAccessor('_140062263790704_c2') + | | | +- GuardManager: source=G['_140062263790704_c2'].Tensor, accessed_by=GetAttrGuardAccessor(Tensor) + | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c2'].Tensor, 82181376) + | | | | +- GuardManager: source=G['_140062263790704_c2'].Tensor.__bases__, accessed_by=GetAttrGuardAccessor(__bases__) + | | | | | +- GuardManager: source=G['_140062263790704_c2'].Tensor.__bases__[0], accessed_by=TupleGetItemGuardAccessor(0) + | | | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c2'].Tensor.__bases__[0], 140062119703136) + | | +- GuardManager: source=G['__builtins_dict___4'], accessed_by=DictGetItemGuardAccessor('__builtins_dict___4') + | | | +- GuardManager: source=G['__builtins_dict___4']['len'], accessed_by=DictGetItemGuardAccessor('len') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['len'], 140062269592480) + | | | +- GuardManager: source=G['__builtins_dict___4']['sum'], accessed_by=DictGetItemGuardAccessor('sum') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['sum'], 140062269593600) + | | | +- GuardManager: source=G['__builtins_dict___4']['list'], accessed_by=DictGetItemGuardAccessor('list') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['list'], 8841312) + | | | +- GuardManager: source=G['__builtins_dict___4']['type'], accessed_by=DictGetItemGuardAccessor('type') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['type'], 8810240) + | | | +- GuardManager: source=G['__builtins_dict___4']['tuple'], accessed_by=DictGetItemGuardAccessor('tuple') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['tuple'], 8812224) + | | | +- GuardManager: source=G['__builtins_dict___4']['object'], accessed_by=DictGetItemGuardAccessor('object') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['object'], 8810976) + | | | +- GuardManager: source=G['__builtins_dict___4']['isinstance'], accessed_by=DictGetItemGuardAccessor('isinstance') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___4']['isinstance'], 140062269592160) + | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_utils_dot__pytree') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140057600329744) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].TreeSpec, accessed_by=GetAttrGuardAccessor(TreeSpec) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 84866496) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 86791872) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf, accessed_by=GetAttrGuardAccessor(_is_leaf) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140411217262720) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140057601749808) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, accessed_by=GetAttrGuardAccessor(_LEAF_SPEC) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 85171104) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 87025712) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, accessed_by=GetAttrGuardAccessor(type) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, accessed_by=GetAttrGuardAccessor(context) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes, accessed_by=GetAttrGuardAccessor(num_nodes) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes == 1 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_leaves, accessed_by=GetAttrGuardAccessor(num_leaves) @@ -2228,116 +2908,135 @@ V1003 10:11:02.818000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children, accessed_by=GetAttrGuardAccessor(num_children) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children == 0 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, accessed_by=GetAttrGuardAccessor(children_specs) - | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8844320) + | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8841312) | | | | | +- LENGTH_CHECK: not G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type, accessed_by=GetAttrGuardAccessor(_get_node_type) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140411217262448) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140057601749536) | | | +- DictGuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES, accessed_by=GetAttrGuardAccessor(SUPPORTED_NODES) - | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 519596 + | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 366913 | | | | +- KeyValueManager pair at index=1 | | | | | +- ValueManager: GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]] | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn, accessed_by=GetAttrGuardAccessor(flatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140411196281984) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140057599989696) | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn, accessed_by=GetAttrGuardAccessor(unflatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140411217182288) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140057600267792) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper, accessed_by=GetAttrGuardAccessor(_tree_flatten_helper) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140411217413040) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140057600001392) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance, accessed_by=GetAttrGuardAccessor(_is_namedtuple_instance) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140411217412592) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140057600000944) -V1003 10:11:02.818000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "0a198eacfa70f28e771164e9f21c0377"} +V1206 15:24:50.370000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "5d9cf4ebb554cb0afa821efb370971a5"} { "name": "entire_frame_compile", - "ts": 1727975462818771.2, + "ts": 1733527490370919.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.819000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "fe22d36e59a389d50249ec2f0d81445a"} +V1206 15:24:50.373000 1667746 torch/_dynamo/utils.py:1045] {"compilation_metrics": {"aot_autograd_cumulative_compile_time_us": 276, "dynamo_cumulative_compile_time_us": 88329, "frame_key": "1", "co_name": "_flex_attention_hop_wrapper", "co_filename": "/data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py", "co_firstlineno": 1303, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 58, "shape_env_guard_count": 0, "graph_op_count": 3, "graph_node_count": 17, "graph_input_count": 11, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": false, "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 9223372036854775807, \"accumulated_cache_size_limit\": 9223372036854775807, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}", "is_forward": true, "dynamo_compile_time_before_restart_us": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.373000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "0e7b5fdc5a2525dc764f709284ebcfb6"} { - "name": "_compile.compile_inner", - "ts": 1727975462819091.0, + "name": "dynamo", + "ts": 1733527490373678.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 0, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "0/0", + "frame_key": "1", + "co_name": "_flex_attention_hop_wrapper", + "co_filename": "/data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py", + "co_firstlineno": 1303, + "cache_size": 0, + "accumulated_cache_size": 0, + "guard_count": 58, + "shape_env_guard_count": 0, + "graph_op_count": 3, + "graph_node_count": 17, + "graph_input_count": 11, + "fail_type": null, + "fail_reason": null, + "fail_user_frame_filename": null, + "fail_user_frame_lineno": null, + "non_compliant_ops": [], + "compliant_custom_ops": [], + "restart_reasons": [], + "dynamo_time_before_restart_s": 0.0, + "has_guarded_code": true, + "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 9223372036854775807, \"accumulated_cache_size_limit\": 9223372036854775807, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.819000 2235078 torch/_dynamo/utils.py:840] {"compilation_metrics": {"compile_id": "0/0", "frame_key": "1", "co_name": "_flex_attention_hop_wrapper", "co_filename": "/data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py", "co_firstlineno": 1049, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 55, "shape_env_guard_count": 0, "graph_op_count": 12, "graph_node_count": 26, "graph_input_count": 11, "start_time": 1727975462.6493652, "entire_frame_compile_time_s": 0.16928768157958984, "backend_compile_time_s": 0.0006258487701416016, "inductor_compile_time_s": null, "code_gen_time_s": null, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "possibly_missed_reinplacing_opportunities": 0, "remote_cache_time_saved_s": 0, "structured_logging_overhead_s": 0.032939939, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": true}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.823000 2235078 torch/_dynamo/convert_frame.py:915] {"dynamo_start": {"stack": [{"line": 916, "name": "", "filename": 1}, {"line": 14, "name": "run_tests", "filename": 2}, {"line": 38, "name": "run_tests", "filename": 3}, {"line": 1273, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3116, "name": "run", "filename": 4}, {"line": 3088, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 2983, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 405, "name": "test_flex_attention_caching", "filename": 1}, {"line": 379, "name": "fn", "filename": 1}]}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.824000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "5c328b3ed733b4526b0979bdb2929653"} +V1206 15:24:50.375000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "1f4a76ca6b5fcae61cfcc92045397262"} { - "name": "_compile.compile_inner", - "ts": 1727975462824011.2, - "args": null, + "name": "dynamo", + "ts": 1733527490375674.2, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.824000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "7333e4a8a9e70773a0188ced97e391ef"} +V1206 15:24:50.376000 1667746 torch/_dynamo/convert_frame.py:961] {"dynamo_start": {"stack": [{"line": 1439, "name": "", "filename": 1}, {"line": 15, "name": "run_tests", "filename": 2}, {"line": 39, "name": "run_tests", "filename": 3}, {"line": 1353, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3234, "name": "run", "filename": 4}, {"line": 3206, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 3099, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 483, "name": "test_flex_attention_caching", "filename": 1}, {"line": 459, "name": "fn", "filename": 1}]}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.376000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "8d1bd9bb5e587d5f1fa509e316288846"} { "name": "entire_frame_compile", - "ts": 1727975462824011.2, - "args": null, + "ts": 1733527490376205.0, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:02.826000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 256, "size": 524288}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.827000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.827000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 0, "source": "L['q']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.836000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 1, "describer_id": 256, "size": 524288}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.836000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 1, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.836000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 1, "source": "L['k']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.838000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 2, "describer_id": 256, "size": 524288}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.838000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 2, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.838000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 2, "source": "L['v']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.847000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 3, "describer_id": 256, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.847000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 3, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.847000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 3, "source": "L['block_mask'].kv_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.985000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 4, "describer_id": 256, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.985000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.985000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 7, "source": "L['block_mask'].kv_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.987000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 5, "describer_id": 256, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.987000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.987000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 8, "source": "L['block_mask'].full_kv_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.989000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 6, "describer_id": 256, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.989000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.989000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 9, "source": "L['block_mask'].full_kv_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.991000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 7, "describer_id": 256, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.991000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.991000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 10, "source": "L['block_mask'].q_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.992000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 8, "describer_id": 256, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.993000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.993000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 11, "source": "L['block_mask'].q_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.994000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 9, "describer_id": 256, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.995000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 12, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.995000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 12, "source": "L['block_mask'].full_q_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.996000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 10, "describer_id": 256, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.997000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 13, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 256}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:02.997000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 256, "id": 13, "source": "L['block_mask'].full_q_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.005000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_output_graph": {"sizes": {"l_q_": [1, 4, 512, 64], "l_k_": [1, 4, 512, 64], "l_v_": [1, 4, 512, 64], "l_block_mask_kv_num_blocks": [1, 1, 16], "l_block_mask_kv_indices": [1, 1, 16, 16], "l_block_mask_full_kv_num_blocks": [1, 1, 16], "l_block_mask_full_kv_indices": [1, 1, 16, 16], "l_block_mask_q_num_blocks": [1, 1, 16], "l_block_mask_q_indices": [1, 1, 16, 16], "l_block_mask_full_q_num_blocks": [1, 1, 16], "l_block_mask_full_q_indices": [1, 1, 16, 16], "child_1": [], "child_2": [], "child_3": [], "child_4": [], "child": [], "child_5": [], "child_6": [], "child_7": [], "child_8": [], "out": [1, 4, 512, 64]}}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0365bd76b3474fb6d96e4c3c42585fb7"} +V1206 15:24:50.377000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 0, "describer_id": 245, "size": 524288}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.377000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.377000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 0, "source": "L['q']"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.383000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 1, "describer_id": 245, "size": 524288}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.384000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 1, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.384000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 1, "source": "L['k']"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.384000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 2, "describer_id": 245, "size": 524288}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.385000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 2, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.385000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 2, "source": "L['v']"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.390000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 3, "describer_id": 245, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.390000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 3, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.390000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 3, "source": "L['block_mask'].kv_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.457000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 4, "describer_id": 245, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.457000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.458000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 7, "source": "L['block_mask'].kv_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.458000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 5, "describer_id": 245, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.459000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.459000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 8, "source": "L['block_mask'].full_kv_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.459000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 6, "describer_id": 245, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.460000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.460000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 9, "source": "L['block_mask'].full_kv_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.460000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 7, "describer_id": 245, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.461000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.461000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 10, "source": "L['block_mask'].q_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.461000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 8, "describer_id": 245, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.462000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.462000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 11, "source": "L['block_mask'].q_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.462000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 9, "describer_id": 245, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.463000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 12, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.463000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 12, "source": "L['block_mask'].full_q_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.463000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 10, "describer_id": 245, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.464000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 13, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 245}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.464000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 245, "id": 13, "source": "L['block_mask'].full_q_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.468000 1667746 torch/_dynamo/output_graph.py:1336] {"dynamo_output_graph": {"sizes": {"l_q_": [1, 4, 512, 64], "l_k_": [1, 4, 512, 64], "l_v_": [1, 4, 512, 64], "l_block_mask_kv_num_blocks": [1, 1, 16], "l_block_mask_kv_indices": [1, 1, 16, 16], "l_block_mask_full_kv_num_blocks": [1, 1, 16], "l_block_mask_full_kv_indices": [1, 1, 16, 16], "l_block_mask_q_num_blocks": [1, 1, 16], "l_block_mask_q_indices": [1, 1, 16, 16], "l_block_mask_full_q_num_blocks": [1, 1, 16], "l_block_mask_full_q_indices": [1, 1, 16, 16], "out": [1, 4, 512, 64]}}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "e6aa2ecb5d533ab181a6215ce16f359c"} class GraphModule(torch.nn.Module): def forward(self, L_q_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_k_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_v_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_block_mask_kv_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_kv_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_block_mask_full_kv_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_full_kv_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_block_mask_q_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_q_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_block_mask_full_q_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_full_q_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): l_q_ = L_q_ @@ -2352,119 +3051,364 @@ V1003 10:11:03.005000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_outpu l_block_mask_full_q_num_blocks = L_block_mask_full_q_num_blocks l_block_mask_full_q_indices = L_block_mask_full_q_indices - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( - child_1: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_1 = None - child_2: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_2 = None - child_3: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_3 = None - child_4: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_4 = None - child: "f32[][]cuda:0" = l_q_.new_empty([], requires_grad = False); child = None + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( score_mod_0 = self.score_mod_0 - child_5: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_5 = None - child_6: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_6 = None - child_7: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_7 = None - child_8: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_8 = None mask_fn_0 = self.mask_fn_0 - flex_attention = torch.ops.higher_order.flex_attention(l_q_, l_k_, l_v_, score_mod_0, (l_block_mask_kv_num_blocks, l_block_mask_kv_indices, l_block_mask_full_kv_num_blocks, l_block_mask_full_kv_indices, l_block_mask_q_num_blocks, l_block_mask_q_indices, l_block_mask_full_q_num_blocks, l_block_mask_full_q_indices, 128, 128, mask_fn_0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); l_q_ = l_k_ = l_v_ = score_mod_0 = l_block_mask_kv_num_blocks = l_block_mask_kv_indices = l_block_mask_full_kv_num_blocks = l_block_mask_full_kv_indices = l_block_mask_q_num_blocks = l_block_mask_q_indices = l_block_mask_full_q_num_blocks = l_block_mask_full_q_indices = mask_fn_0 = None + flex_attention = torch.ops.higher_order.flex_attention(l_q_, l_k_, l_v_, score_mod_0, (l_block_mask_kv_num_blocks, l_block_mask_kv_indices, l_block_mask_full_kv_num_blocks, l_block_mask_full_kv_indices, l_block_mask_q_num_blocks, l_block_mask_q_indices, l_block_mask_full_q_num_blocks, l_block_mask_full_q_indices, 128, 128, mask_fn_0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); l_q_ = l_k_ = l_v_ = score_mod_0 = l_block_mask_kv_num_blocks = l_block_mask_kv_indices = l_block_mask_full_kv_num_blocks = l_block_mask_full_kv_indices = l_block_mask_q_num_blocks = l_block_mask_q_indices = l_block_mask_full_q_num_blocks = l_block_mask_full_q_indices = mask_fn_0 = None out: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0]; flex_attention = None return (out,) class score_mod_0(torch.nn.Module): def forward(self, child: "f32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0", child_4: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:377 in score_mod, code: return score + (q - kv) + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:457 in score_mod, code: return score + (q - kv) sub: "i32[][]cuda:0" = child_3 - child_4; child_3 = child_4 = None add: "f32[][]cuda:0" = child + sub; child = sub = None return add class mask_fn_0(torch.nn.Module): - def forward(self, child_5: "i32[][]cuda:0", child_6: "i32[][]cuda:0", child_7: "i32[][]cuda:0", child_8: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 - ge: "b8[][]cuda:0" = child_7 >= child_8; child_7 = child_8 = None + def forward(self, child: "i32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0"): + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + ge: "b8[][]cuda:0" = child_2 >= child_3; child_2 = child_3 = None return ge -V1003 10:11:03.006000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d0fdf0bbe0d39c70da0305c36b970c1c"} +V1206 15:24:50.468000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "80655624a23088893ab4a2c409ac95a3"} { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975463006504.2, - "args": null, + "name": "backend_compile", + "ts": 1733527490468948.0, + "args": { + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.006000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "2c12bb411b8d14ec0cf2ffdf2c4427c3"} +V1206 15:24:50.469000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "a719c6c37340b85f41f898361ba8c722"} { - "name": "backend_compile", - "ts": 1727975463006504.2, - "args": null, + "name": "_recursive_pre_grad_passes", + "ts": 1733527490469220.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.012000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "17a34747abc18cf69e8e2a0141a3ab09"} +V1206 15:24:50.469000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "a15be333d609301edbc3c4ba97a6e27a"} + { + "name": "_recursive_pre_grad_passes", + "ts": 1733527490469741.8, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.471000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c4c11179a8b2ec9ad87664f12d76dd56"} { "name": "create_aot_dispatcher_function", - "ts": 1727975463012509.8, - "args": null, + "ts": 1733527490471944.8, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.138000 2235078 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:215] {"aot_forward_graph": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f6cec159cb1cee52405afcfe40b62c03"} +V1206 15:24:50.519000 1667746 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:214] {"artifact": {"name": "aot_forward_graph_fw_metadata", "encoding": "string"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d952b3071dd9ba606a04d644841d9aba"} + ViewAndMutationMeta(input_info=[InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True)], + output_info=[OutputAliasInfo(output_type=, + raw_type=, + base_idx=None, + dynamic_dims=set(), + requires_grad=False, + functional_tensor=None)], + num_intermediate_bases=0, + keep_input_mutations=True, + traced_tangents=[], + subclass_inp_meta=[PlainTensorMeta(unwrapped_idx=0, + memory_format=None), + PlainTensorMeta(unwrapped_idx=1, + memory_format=None), + PlainTensorMeta(unwrapped_idx=2, + memory_format=None), + PlainTensorMeta(unwrapped_idx=3, + memory_format=None), + PlainTensorMeta(unwrapped_idx=4, + memory_format=None), + PlainTensorMeta(unwrapped_idx=5, + memory_format=None), + PlainTensorMeta(unwrapped_idx=6, + memory_format=None), + PlainTensorMeta(unwrapped_idx=7, + memory_format=None), + PlainTensorMeta(unwrapped_idx=8, + memory_format=None), + PlainTensorMeta(unwrapped_idx=9, + memory_format=None), + PlainTensorMeta(unwrapped_idx=10, + memory_format=None)], + subclass_fw_graph_out_meta=[PlainTensorMeta(unwrapped_idx=0, + memory_format=None)], + subclass_tangent_meta=[], + is_train=False, + traced_tangent_metas=None, + num_symints_saved_for_bw=None, + grad_enabled_mutation=None, + deterministic=None, + static_input_indices=[], + tokens={}, + indices_of_inputs_that_requires_grad_with_mutations_in_bw=[], + bw_donated_idxs=None, + num_backward_tokens=0) +V1206 15:24:50.520000 1667746 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:232] {"aot_inference_graph": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "02884732a6f03bac463b0e1e5f9b1cf4"} class (torch.nn.Module): def forward(self, arg0_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg1_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg2_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg3_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg4_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg5_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg6_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg7_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg8_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg9_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg10_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( sdpa_score0 = self.sdpa_score0 sdpa_mask0 = self.sdpa_mask0 - flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None + flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None getitem: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0]; flex_attention = None return (getitem,) class sdpa_score0(torch.nn.Module): def forward(self, arg0_1: "f32[][]cuda:0", arg1_1: "i32[][]cuda:0", arg2_1: "i32[][]cuda:0", arg3_1: "i32[][]cuda:0", arg4_1: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( sub: "i32[][]cuda:0" = torch.ops.aten.sub.Tensor(arg3_1, arg4_1); arg3_1 = arg4_1 = None add: "f32[][]cuda:0" = torch.ops.aten.add.Tensor(arg0_1, sub); arg0_1 = sub = None return add class sdpa_mask0(torch.nn.Module): def forward(self, arg0_1: "i32[][]cuda:0", arg1_1: "i32[][]cuda:0", arg2_1: "i32[][]cuda:0", arg3_1: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 ge: "b8[][]cuda:0" = torch.ops.aten.ge.Tensor(arg2_1, arg3_1); arg2_1 = arg3_1 = None return ge -V1003 10:11:03.139000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "96a6ba3562cc8fdc9267e57d6099f6dd"} +V1206 15:24:50.521000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "2386afea16835c738f6306784edc87e8"} { "name": "compile_fx..fw_compiler_base", - "ts": 1727975463139343.8, - "args": null, + "ts": 1733527490521683.8, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.141000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "4405e9a49d13d3764e0174322946bd85"} +V1206 15:24:50.522000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "76ae4cce2f52ad7feca705fd70cb3c6c"} { - "name": "compile_fx_inner", - "ts": 1727975463141246.5, - "args": null, + "name": "_recursive_joint_graph_passes", + "ts": 1733527490522094.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.141000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "98ce9c89d4855d24e864b2a507237295"} +V1206 15:24:50.523000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c0a313f319086d99c7003f86a39e8eb0"} + { + "name": "_recursive_joint_graph_passes", + "ts": 1733527490523142.8, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.523000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "51c50f90b4f3e9efaa73f45c2709dc93"} { "name": "inductor_compile", - "ts": 1727975463141246.5, - "args": null, + "ts": 1733527490523375.8, + "args": { + "fn_name": "compile_fx_inner", + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.531000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "30b00efd4f6faf8b068719e99390e687"} + { + "name": "TritonBundler.read_and_emit", + "ts": 1733527490531851.0, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_output_code": {"filename": "/tmp/oulgen/tmp4z1i5ywe/kp/ckpysuucou6gm55terbvpynnfevubjpwkfm3ubzxyauw2bgggpvi.py"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "187381168c614ba1181a424c67fdb77c"} +V1206 15:24:50.532000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d06401ea7dbeea897c26573b7848e594"} + { + "name": "TritonBundler.read_and_emit", + "ts": 1733527490532212.2, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.532000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "e60ce588f8a8da2d850b98d827e6543e"} + { + "name": "PyCodeCache.load_by_key_path", + "ts": 1733527490532650.8, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.545000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d4c6869eeaa9ef14de6b518a45d9c651"} + { + "name": "async_compile.wait", + "ts": 1733527490545938.2, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.599000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "a05612636136909ead93ebf04e1a42ad"} + { + "name": "async_compile.wait", + "ts": 1733527490599093.0, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.599000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "eb0838b082ff5e829a02f8ce1aae27cf"} + { + "name": "PyCodeCache.load_by_key_path", + "ts": 1733527490599434.2, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.599000 1667746 torch/_inductor/codecache.py:1267] {"inductor_output_code": {"filename": "/tmp/tmpsn77tlsm/sm/csmb7f23sz2oa7mngdcoprpgrrbziqzirwhkjffdmxv6qhdf3e57.py"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "9ced76d2b85dd84790793e85fcadcc28"} # AOT ID: ['0_inference'] from ctypes import c_void_p, c_long, c_int import torch @@ -2484,7 +3428,14 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp from torch._C import _cuda_getCurrentRawStream as get_raw_stream import triton import triton.language as tl - from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph + from torch._inductor.runtime.triton_heuristics import ( + grid, + split_scan_grid, + grid_combo_kernels, + start_graph, + end_graph, + cooperative_reduction_grid, + ) aten = torch.ops.aten inductor_ops = torch.ops.inductor @@ -2496,14 +3447,15 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor alloc_from_pool = torch.ops.inductor._alloc_from_pool async_compile = AsyncCompile() + empty_strided_p2p = torch._C._distributed_c10d._SymmetricMemory.empty_strided_p2p - # kernel path: /tmp/oulgen/tmp4z1i5ywe/ke/ckedjitfju7kpaxnpatsaz3gzkdz6znc4u5v2e22w2xzbjdvecy5.py + # kernel path: /tmp/tmpsn77tlsm/xo/cxokkamvqulf4elvqs25iez7wbmqqfbp3i4qy7g2hwvher2zrziz.py # Topologically Sorted Source Nodes: [flex_attention], Original ATen: [] # Source node to ATen node mapping: # flex_attention => flex_attention # Graph fragment: - # %flex_attention : [num_users=1] = call_function[target=torch.ops.higher_order.flex_attention](args = (%arg0_1, %arg1_1, %arg2_1, %sdpa_score0, (%arg3_1, %arg4_1, %arg5_1, %arg6_1, %arg7_1, %arg8_1, %arg9_1, %arg10_1, 128, 128, %sdpa_mask0), 0.125, {ROWS_GUARANTEED_SAFE: False, PRESCALE_QK: False, OUTPUT_LOGSUMEXP: False}, (), ()), kwargs = {}) + # %flex_attention : [num_users=1] = call_function[target=torch.ops.higher_order.flex_attention](args = (%arg0_1, %arg1_1, %arg2_1, %sdpa_score0, (%arg3_1, %arg4_1, %arg5_1, %arg6_1, %arg7_1, %arg8_1, %arg9_1, %arg10_1, 128, 128, %sdpa_mask0), 0.125, {PRESCALE_QK: False, ROWS_GUARANTEED_SAFE: False, BLOCKS_ARE_CONTIGUOUS: False, OUTPUT_LOGSUMEXP: True}, (), ()), kwargs = {}) triton_tem_fused_0 = async_compile.triton('triton_tem_fused_0', ''' import triton import triton.language as tl @@ -2511,19 +3463,20 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp from torch._inductor.runtime import triton_helpers, triton_heuristics from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math - from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties + from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, DeviceProperties @triton_heuristics.template( num_stages=3, num_warps=4, - triton_meta={'signature': {'arg_Q': '*fp32', 'arg_K': '*fp32', 'arg_V': '*fp32', 'arg_LSE': '*fp32', 'arg_KV_NUM_BLKS': '*i32', 'arg_KV_IDX': '*i32', 'arg_FULL_KV_NUM_BLKS': '*i32', 'arg_FULL_KV_IDX': '*i32', 'out_ptr0': '*fp32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108, warp_size=32), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]}, - inductor_meta={'kernel_name': 'triton_tem_fused_0', 'backend_hash': 'FB2CA426CF35F271C56C0D69873498391AC248E25890F2B631CA8B52D56952BD', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': False, 'autotune_pointwise': True, 'autotune_remote_cache': False, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}, + triton_meta={'signature': {'arg_Q': '*fp32', 'arg_K': '*fp32', 'arg_V': '*fp32', 'arg_LSE': '*fp32', 'arg_KV_NUM_BLKS': '*i32', 'arg_KV_IDX': '*i32', 'arg_FULL_KV_NUM_BLKS': '*i32', 'arg_FULL_KV_IDX': '*i32', 'out_ptr0': '*fp32'}, 'device': DeviceProperties(type='cuda', index=0, cc=90, major=9, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=132, warp_size=32), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]}, + inductor_meta={'kernel_name': 'triton_tem_fused_0', 'backend_hash': '562E840D41CEB1D8E51DE726EA7592B0C37A0C6FBD72CF4E958863CEC11D41A7', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': False, 'autotune_pointwise': True, 'autotune_remote_cache': False, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}, ) @triton.jit def triton_tem_fused_0(arg_Q, arg_K, arg_V, arg_LSE, arg_KV_NUM_BLKS, arg_KV_IDX, arg_FULL_KV_NUM_BLKS, arg_FULL_KV_IDX, out_ptr0): - ROWS_GUARANTEED_SAFE : tl.constexpr = False PRESCALE_QK : tl.constexpr = False - OUTPUT_LOGSUMEXP : tl.constexpr = False + ROWS_GUARANTEED_SAFE : tl.constexpr = False + BLOCKS_ARE_CONTIGUOUS : tl.constexpr = False + OUTPUT_LOGSUMEXP : tl.constexpr = True FLOAT32_PRECISION : tl.constexpr = 'ieee' IS_DIVISIBLE : tl.constexpr = True SM_SCALE : tl.constexpr = 0.125 @@ -2571,6 +3524,8 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp # about 20% more numerical error, but slightly faster. # ROWS_GUARANTEED_SAFE: Is it guaranteed that at least one value in each row # is not masked out? If so, we can skip an extra safety check + # BLOCKS_ARE_CONTIGUOUS: Is it guaranteed that all blocks in the mask are + # contiguous? If so, we don't need to do an indirect jump for every block tl.static_assert(SPARSE_Q_BLOCK_SIZE >= BLOCK_M and SPARSE_Q_BLOCK_SIZE % BLOCK_M == 0) tl.static_assert(SPARSE_KV_BLOCK_SIZE >= BLOCK_N and SPARSE_KV_BLOCK_SIZE % BLOCK_N == 0) @@ -2736,11 +3691,10 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp idx_d = tl.arange(0, V_HEAD_DIM)[None, :] mask = idx_m < Q_LEN - # TODO generalize and add proper mask support - xindex = idx_d + (64*idx_m) + (32768*idx_hq) + (131072*idx_zq) - tl.store(out_ptr0 + (tl.broadcast_to(idx_d + (64*idx_m) + (32768*idx_hq), acc.shape)), acc, mask) - # TODO dont want to write this if we dont require grad + xindex = idx_d + 64*idx_m + 32768*idx_hq + 131072*idx_zq + tl.store(out_ptr0 + (tl.broadcast_to(idx_d + 64*idx_m + 32768*idx_hq, acc.shape)), acc, mask) + if OUTPUT_LOGSUMEXP: off_hz = tl.program_id(1) l_ptrs = LSE + off_hz * Q_LEN + offs_m @@ -2767,9 +3721,10 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp IS_FULL_BLOCKS, ): # Redefines all kernel parameters (BLOCK_M, etc.) so we don't need to plumb them all through - ROWS_GUARANTEED_SAFE : tl.constexpr = False PRESCALE_QK : tl.constexpr = False - OUTPUT_LOGSUMEXP : tl.constexpr = False + ROWS_GUARANTEED_SAFE : tl.constexpr = False + BLOCKS_ARE_CONTIGUOUS : tl.constexpr = False + OUTPUT_LOGSUMEXP : tl.constexpr = True FLOAT32_PRECISION : tl.constexpr = 'ieee' IS_DIVISIBLE : tl.constexpr = True SM_SCALE : tl.constexpr = 0.125 @@ -2821,7 +3776,7 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp # update pointers offset = get_offset_for_next_block( start_n, kv_indices, kv_num_blocks, - SPARSE_KV_BLOCK_SIZE, SPARSE_KV_MULTIPLE, BLOCK_N + SPARSE_KV_BLOCK_SIZE, SPARSE_KV_MULTIPLE, BLOCK_N, BLOCKS_ARE_CONTIGUOUS ) V_block_ptr = tl.advance(V_block_ptr, (offset, 0)) @@ -2833,13 +3788,18 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp @triton.jit - def get_offset_for_next_block(loop_iter, col_indices, total_blocks, SPARSE_BLOCK, SPARSE_BLOCK_MULTIPLE, BLOCK): + def get_offset_for_next_block( + loop_iter, col_indices, total_blocks, + SPARSE_BLOCK, SPARSE_BLOCK_MULTIPLE, BLOCK, + BLOCKS_ARE_CONTIGUOUS: tl.constexpr + ): + if BLOCKS_ARE_CONTIGUOUS: + return BLOCK cur_block_idx = loop_iter // SPARSE_BLOCK_MULTIPLE cur_block = tl.load(col_indices + cur_block_idx, eviction_policy="evict_last") next_block = tl.load(col_indices + cur_block_idx + 1, eviction_policy="evict_last", mask=cur_block_idx + 1 < total_blocks) needs_jump = (loop_iter + 1) % SPARSE_BLOCK_MULTIPLE == 0 jump_to_block = (next_block - cur_block ) * SPARSE_BLOCK - (SPARSE_BLOCK_MULTIPLE - 1) * BLOCK - offset = jump_to_block * needs_jump + (1 - needs_jump) * BLOCK return offset @@ -2855,9 +3815,10 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp IS_FULL_BLOCKS, CHECK_BLOCK_BOUNDARY=False, ): # Redefines all kernel parameters (BLOCK_M, etc.) so we don't need to plumb them all through - ROWS_GUARANTEED_SAFE : tl.constexpr = False PRESCALE_QK : tl.constexpr = False - OUTPUT_LOGSUMEXP : tl.constexpr = False + ROWS_GUARANTEED_SAFE : tl.constexpr = False + BLOCKS_ARE_CONTIGUOUS : tl.constexpr = False + OUTPUT_LOGSUMEXP : tl.constexpr = True FLOAT32_PRECISION : tl.constexpr = 'ieee' IS_DIVISIBLE : tl.constexpr = True SM_SCALE : tl.constexpr = 0.125 @@ -2907,7 +3868,7 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp if CHECK_BLOCK_BOUNDARY: - mask_mod_output = tl.where(offs_n < KV_LEN, mask_mod_output, float("-inf")) + mask_mod_output = tl.where(offs_n < KV_LEN, mask_mod_output, False) # apply mask for partially unmasked blocks post_mod_scores = tl.where(mask_mod_output, post_mod_scores, float("-inf")) @@ -2945,7 +3906,7 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp return acc, l_i, m_i ''', device_str='cuda') - meta0 = {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False, 'FLOAT32_PRECISION': "'ieee'", 'IS_DIVISIBLE': True, 'SM_SCALE': 0.125, 'GQA_SHARED_HEADS': 1, 'HAS_FULL_BLOCKS': True, 'QK_HEAD_DIM': 64, 'V_HEAD_DIM': 64, 'BLOCK_M': 128, 'BLOCK_N': 32, 'SPARSE_Q_BLOCK_SIZE': 128, 'SPARSE_KV_BLOCK_SIZE': 128} + meta0 = {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True, 'FLOAT32_PRECISION': "'ieee'", 'IS_DIVISIBLE': True, 'SM_SCALE': 0.125, 'GQA_SHARED_HEADS': 1, 'HAS_FULL_BLOCKS': True, 'QK_HEAD_DIM': 64, 'V_HEAD_DIM': 64, 'BLOCK_M': 128, 'BLOCK_N': 32, 'SPARSE_Q_BLOCK_SIZE': 128, 'SPARSE_KV_BLOCK_SIZE': 128} async_compile.wait(globals()) @@ -3005,14 +3966,15 @@ V1003 10:11:03.282000 2235078 torch/_inductor/codecache.py:1138] {"inductor_outp from torch._inductor.wrapper_benchmark import compiled_module_main compiled_module_main('None', benchmark_compiled_module) -V1003 10:11:03.283000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "55abfbf07c349adf63e8871178bf974b"} +V1206 15:24:50.600000 1667746 torch/_dynamo/utils.py:1327] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0b77d8bc6ee36d773c74713ef29c7921"} { "name": "fx_graph_cache_hit", - "ts": 1727975463283186.5, + "ts": 1733527490524020.8, "args": { - "key": "f4lkea5y7lzhlshohvr3aqpd7bchdflfs7j5wn7mrurponawoutk", + "triton_bundler_meta": "TritonBundlerMetadata(cached_kernel_names=[])", + "key": "ff6i4wr5gqsaog5zgs3qrwebcodgoghruloxh5pzs6iul3cmayau", "components": [ - "[n7x23yy6fih6vdcjzlzbhy3d6vx3ilu7ylp3zz6wkultu4yvnzn] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[v4wl2w2eawehu2yuzv4ftucrovobmkotky2oimjdvonwbz3g4ir] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", @@ -3025,6 +3987,7 @@ V1003 10:11:03.283000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", + "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", @@ -3032,7 +3995,6 @@ V1003 10:11:03.283000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", - "[f44ag5aflby2bkxl7a4k6whljrk7jat7bmreuxklei4p3czhk7p] fx_kwargs[user_visible_outputs]: {'getitem': None}", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", @@ -3046,399 +4008,701 @@ V1003 10:11:03.283000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", - "[7uhqwjfn75ek3woo3k7em2mluon5hx2ojvzlevlvjvz6xfxjhzl] torch_version: ", - "[c3z7bmoxyo6gl5hi47v6dc7jwsl55b3asd75nr25uyengi5ah3p] system_info[device]: {'name': 'NVIDIA PG509-210'}", - "[3fb7kae6ogkdd4zcm3fkjoipdpybxhn4aoxzv7z7xsfwq233e4l] system_info[version]: {'triton': '3.1.0+5fe38ffd73dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-a8fb7be728d460b7ec64ab62edb8af1bbca8994fd718cde7178d46bad64530a1-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.0'}", - "[z5x2bdhir5lzlbti73vdbfulnuu5vinzpwgmmgf4rjb775tzl3h] system_info[hash]: 9698c97edde4a99a2f3b54bbd0db5291bbcdb75c83acb376ccff61fb0bf0ac1a", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[abi_compatible]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_dump_consts_bin]: False", - "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", - "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", - "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", - "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[bundled_autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", + "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", + "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", + "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", + "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", - "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", + "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", - "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", - "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", - "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", - "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cuda.generate_test_runner]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", - "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" + ], + "cache_event_time": 1733527490600011529, + "cache_state": "hit", + "time_saved_ns": 4009141776, + "compile_id": "1/0" + }, + "ph": "i", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0, + "s": "p" + } +V1206 15:24:50.600000 1667746 torch/_inductor/compile_fx.py:751] {"artifact": {"name": "fx_graph_cache_hit", "encoding": "json"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d8cb8a6a0ac5b1d58d28337f3f1a055a"} + {"triton_bundler_meta": "TritonBundlerMetadata(cached_kernel_names=[])", "key": "ff6i4wr5gqsaog5zgs3qrwebcodgoghruloxh5pzs6iul3cmayau", "components": ["[v4wl2w2eawehu2yuzv4ftucrovobmkotky2oimjdvonwbz3g4ir] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[3]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[4]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[5]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[6]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[7]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[8]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", "[kcuxe2zwm3mzv2uk6adm6iskoy35bqfv725twacrdewod2dbl5d] inputs_to_check[3]: 3", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inputs_to_check[4]: 4", "[qs5hilycp4ew4ivtc7m5jaxp7q4pm5slioxw3fi3ur6ei65ybz4] inputs_to_check[5]: 5", "[agkvbkaha53nbz3aeeuhvxjvvc4glhfjofzkg6g2qjoo2e5otcx] inputs_to_check[6]: 6", "[j3s5elu6itwgjafc7rzhy4whrbufl6kfmlufjhh25grt643bk5f] inputs_to_check[7]: 7", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inputs_to_check[8]: 8", "[qlgfiyqewrmkgqth2qm6wkq2ja5lzkapg3ypgnvoyfqqnidaoj3] inputs_to_check[9]: 9", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[bundled_autotune_remote_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None"], "cache_event_time": 1733527490600011529, "cache_state": "hit", "time_saved_ns": 4009141776, "compile_id": "1/0"} +V1206 15:24:50.601000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "13cc116eca5abf2a6d7e2ad12c7d39c5"} + { + "name": "inductor_compile", + "ts": 1733527490600831.8, + "args": { + "fn_name": "compile_fx_inner", + "compile_id": "1/0", + "is_backward": false, + "cached_kernel_names": [], + "cache_state": "hit", + "cache_event_time": 1733527490524020812, + "key": "ff6i4wr5gqsaog5zgs3qrwebcodgoghruloxh5pzs6iul3cmayau", + "components": [ + "[v4wl2w2eawehu2yuzv4ftucrovobmkotky2oimjdvonwbz3g4ir] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[3]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[4]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[5]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[6]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[7]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[8]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", + "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", + "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", + "[kcuxe2zwm3mzv2uk6adm6iskoy35bqfv725twacrdewod2dbl5d] inputs_to_check[3]: 3", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inputs_to_check[4]: 4", + "[qs5hilycp4ew4ivtc7m5jaxp7q4pm5slioxw3fi3ur6ei65ybz4] inputs_to_check[5]: 5", + "[agkvbkaha53nbz3aeeuhvxjvvc4glhfjofzkg6g2qjoo2e5otcx] inputs_to_check[6]: 6", + "[j3s5elu6itwgjafc7rzhy4whrbufl6kfmlufjhh25grt643bk5f] inputs_to_check[7]: 7", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inputs_to_check[8]: 8", + "[qlgfiyqewrmkgqth2qm6wkq2ja5lzkapg3ypgnvoyfqqnidaoj3] inputs_to_check[9]: 9", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", + "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", + "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[enable_auto_functionalized_v2]: False", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", - "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", - "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", - "[62lrdx35b7hnumwb7mp5oc5y5csm2abylvtdzfloct3noaqov3n] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'post_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", - "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", - "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", - "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", - "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[bundled_autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", - "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", - "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", - "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", - "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", + "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", - "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", - "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", + "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", - "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", - "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_post_pass]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_pre_pass]: None", - "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", - "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", - "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", - "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", - "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", - "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_peak_memory]: False", - "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", - "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", - "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", - "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", + "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", - "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", - "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", - "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", - "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", - "[vzzema5ityqj2wepdmkulue7q5pcevdr5h27oxxutf35d4tjume] inductor_config[triton_kernel_default_layout_constraint]: flexible_layout", - "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", - "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False" + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", + "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" ], - "cache_event_time": 1727975463283186408, - "cache_state": "hit", - "time_saved_ns": 6372999079 - }, - "ph": "i", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0, - "s": "p" - } -V1003 10:11:03.284000 2235078 torch/_inductor/codecache.py:1463] {"artifact": {"name": "fx_graph_cache_hit", "encoding": "json"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b87f8df3afd49918e4211191e3fc10a5"} - {"key": "f4lkea5y7lzhlshohvr3aqpd7bchdflfs7j5wn7mrurponawoutk", "components": ["[n7x23yy6fih6vdcjzlzbhy3d6vx3ilu7ylp3zz6wkultu4yvnzn] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[3]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[4]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[5]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[6]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[7]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[8]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", "[f44ag5aflby2bkxl7a4k6whljrk7jat7bmreuxklei4p3czhk7p] fx_kwargs[user_visible_outputs]: {'getitem': None}", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", "[kcuxe2zwm3mzv2uk6adm6iskoy35bqfv725twacrdewod2dbl5d] inputs_to_check[3]: 3", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inputs_to_check[4]: 4", "[qs5hilycp4ew4ivtc7m5jaxp7q4pm5slioxw3fi3ur6ei65ybz4] inputs_to_check[5]: 5", "[agkvbkaha53nbz3aeeuhvxjvvc4glhfjofzkg6g2qjoo2e5otcx] inputs_to_check[6]: 6", "[j3s5elu6itwgjafc7rzhy4whrbufl6kfmlufjhh25grt643bk5f] inputs_to_check[7]: 7", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inputs_to_check[8]: 8", "[qlgfiyqewrmkgqth2qm6wkq2ja5lzkapg3ypgnvoyfqqnidaoj3] inputs_to_check[9]: 9", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", "[7uhqwjfn75ek3woo3k7em2mluon5hx2ojvzlevlvjvz6xfxjhzl] torch_version: ", "[c3z7bmoxyo6gl5hi47v6dc7jwsl55b3asd75nr25uyengi5ah3p] system_info[device]: {'name': 'NVIDIA PG509-210'}", "[3fb7kae6ogkdd4zcm3fkjoipdpybxhn4aoxzv7z7xsfwq233e4l] system_info[version]: {'triton': '3.1.0+5fe38ffd73dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-a8fb7be728d460b7ec64ab62edb8af1bbca8994fd718cde7178d46bad64530a1-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.0'}", "[z5x2bdhir5lzlbti73vdbfulnuu5vinzpwgmmgf4rjb775tzl3h] system_info[hash]: 9698c97edde4a99a2f3b54bbd0db5291bbcdb75c83acb376ccff61fb0bf0ac1a", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[abi_compatible]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_dump_consts_bin]: False", "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cuda.generate_test_runner]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[enable_auto_functionalized_v2]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", "[62lrdx35b7hnumwb7mp5oc5y5csm2abylvtdzfloct3noaqov3n] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'post_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_pre_pass]: None", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_peak_memory]: False", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", "[vzzema5ityqj2wepdmkulue7q5pcevdr5h27oxxutf35d4tjume] inductor_config[triton_kernel_default_layout_constraint]: flexible_layout", "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False"], "cache_event_time": 1727975463283186408, "cache_state": "hit", "time_saved_ns": 6372999079} -V1003 10:11:03.284000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d86af378fd420082506385ca5acfaccc"} - { - "name": "inductor_compile", - "ts": 1727975463284819.8, - "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } - }, - "ph": "E", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0 - } -V1003 10:11:03.285000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "45657f0a65e5f7e3f3fcfb73f437d842"} - { - "name": "compile_fx_inner", - "ts": 1727975463285151.0, - "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "cache_bypass_reason": null, + "remote_cache_enabled": false, + "local_cache_enabled": true }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.285000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "12df42838dd8ed7060ce2ffd52760f3c"} +V1206 15:24:50.601000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ad742b9e5576eedb50497df147df1721"} { "name": "compile_fx..fw_compiler_base", - "ts": 1727975463285792.0, + "ts": 1733527490601481.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.289000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "1bd305e520f60c583bf3f70d53bf1988"} +V1206 15:24:50.603000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "cb401f4f44ce40d91fb40593e2e0fab8"} { "name": "create_aot_dispatcher_function", - "ts": 1727975463289199.8, + "ts": 1733527490603290.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.289000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "02f3ea38f17491dd957dd1864f5411e6"} +V1206 15:24:50.603000 1667746 torch/_dynamo/utils.py:1327] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "034a66d21d6ca4969f8613317ccf781b"} { - "name": "backend_compile", - "ts": 1727975463289789.5, + "name": "autograd_cache_bypass", + "ts": 1733527490471931.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "cache_bypass_reason": "Unsupported call_function target flex_attention. \n Function module: torch.ops.higher_order, \nFunction name: flex_attention", + "cache_bypass_hard_exception": false, + "key": null, + "cache_state": "bypass", + "components": [], + "compile_id": "1/0" }, - "ph": "E", + "ph": "i", "cat": "dynamo_timed", "tid": 0, - "pid": 0 + "pid": 0, + "s": "p" } -V1003 10:11:03.290000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "573fd1c41404d17c4da25ab7cfc13f4b"} +V1206 15:24:50.603000 1667746 torch/_functorch/_aot_autograd/autograd_cache.py:763] {"artifact": {"name": "aotautograd_cache_hash", "encoding": "json"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "5601d02186053adcc1ba29fd248c1d20"} + {"cache_bypass_reason": "Unsupported call_function target flex_attention. \n Function module: torch.ops.higher_order, \nFunction name: flex_attention", "cache_bypass_hard_exception": false, "key": null, "cache_state": "bypass", "components": [], "compile_id": "1/0"} +V1206 15:24:50.603000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "bd634b54ab5138da3c38e2434aae7337"} { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975463290072.5, + "name": "backend_compile", + "ts": 1733527490603908.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "1/0", + "requires_subclass_dispatch": false, + "dispatch_mode": "inference", + "cache_state": "bypass", + "cache_event_time": 1733527490471931164, + "key": null, + "components": [], + "cache_bypass_reason": "Unsupported call_function target flex_attention. \n Function module: torch.ops.higher_order, \nFunction name: flex_attention", + "remote_cache_enabled": false, + "local_cache_enabled": true }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.320000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_str": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "6627aec2e890cf84ba5ee985a77950cc"} +V1206 15:24:50.618000 1667746 torch/_dynamo/guards.py:2315] {"dynamo_cpp_guards_str": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "771ab34a6bd9546eb139960f16071592"} TREE_GUARD_MANAGER: +- RootGuardManager - | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:471 in init_ambient_guards + | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:484 in init_ambient_guards | +- GLOBAL_STATE: ___check_global_state() | +- TORCH_FUNCTION_MODE_STACK: ___check_torch_function_mode_stack() - | +- GuardManager: source=L['k'], accessed_by=DictGetItemGuardAccessor(k) - | | +- TYPE_MATCH: ___check_type_id(L['k'], 82028112) + | +- GuardManager: source=L['k'], accessed_by=DictGetItemGuardAccessor('k') + | | +- TYPE_MATCH: ___check_type_id(L['k'], 82181376) | | +- TENSOR_MATCH: check_tensor(L['k'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) | | +- NO_HASATTR: hasattr(L['k'], '_dynamo_dynamic_indices') == False | | +- NO_TENSOR_ALIASING: check_no_aliasing(L['k'], L['q'], L['v'], L['block_mask'].q_indices, L['block_mask'].kv_indices, L['block_mask'].q_num_blocks, L['block_mask'].kv_num_blocks, L['block_mask'].full_q_indices, L['block_mask'].full_kv_indices, L['block_mask'].full_q_num_blocks, L['block_mask'].full_kv_num_blocks) - | +- GuardManager: source=L['q'], accessed_by=DictGetItemGuardAccessor(q) - | | +- TYPE_MATCH: ___check_type_id(L['q'], 82028112) + | | +- GuardManager: source=L['k'].dim, accessed_by=GetAttrGuardAccessor(dim) + | | +- GuardManager: source=L['k'].size, accessed_by=GetAttrGuardAccessor(size) + | +- GuardManager: source=L['q'], accessed_by=DictGetItemGuardAccessor('q') + | | +- TYPE_MATCH: ___check_type_id(L['q'], 82181376) | | +- TENSOR_MATCH: check_tensor(L['q'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) | | +- NO_HASATTR: hasattr(L['q'], '_dynamo_dynamic_indices') == False | | +- NO_TENSOR_ALIASING - | +- GuardManager: source=L['v'], accessed_by=DictGetItemGuardAccessor(v) - | | +- TYPE_MATCH: ___check_type_id(L['v'], 82028112) + | | +- GuardManager: source=L['q'].dim, accessed_by=GetAttrGuardAccessor(dim) + | | +- GuardManager: source=L['q'].size, accessed_by=GetAttrGuardAccessor(size) + | +- GuardManager: source=L['v'], accessed_by=DictGetItemGuardAccessor('v') + | | +- TYPE_MATCH: ___check_type_id(L['v'], 82181376) | | +- TENSOR_MATCH: check_tensor(L['v'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) | | +- NO_HASATTR: hasattr(L['v'], '_dynamo_dynamic_indices') == False | | +- NO_TENSOR_ALIASING - | +- GuardManager: source=L['score_mod'], accessed_by=DictGetItemGuardAccessor(score_mod) + | | +- GuardManager: source=L['v'].dim, accessed_by=GetAttrGuardAccessor(dim) + | | +- GuardManager: source=L['v'].size, accessed_by=GetAttrGuardAccessor(size) + | +- GuardManager: source=L['score_mod'], accessed_by=DictGetItemGuardAccessor('score_mod') | | +- GuardManager: source=L['score_mod'].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | +- ID_MATCH: ___check_obj_id(L['score_mod'].__code__, 140413271879296) - | +- GuardManager: source=L['block_mask'], accessed_by=DictGetItemGuardAccessor(block_mask) - | | +- TYPE_MATCH: ___check_type_id(L['block_mask'], 387600320) + | | | +- ID_MATCH: ___check_obj_id(L['score_mod'].__code__, 140062268556144) + | +- GuardManager: source=L['block_mask'], accessed_by=DictGetItemGuardAccessor('block_mask') + | | +- TYPE_MATCH: ___check_type_id(L['block_mask'], 139679664) | | +- GuardManager: source=L['block_mask'].mask_mod, accessed_by=GetAttrGuardAccessor(mask_mod) | | | +- GuardManager: source=L['block_mask'].mask_mod.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].mask_mod.__code__, 140413271880128) + | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].mask_mod.__code__, 140062268556768) | | +- GuardManager: source=L['block_mask'].q_indices, accessed_by=GetAttrGuardAccessor(q_indices) | | | +- TENSOR_MATCH: check_tensor(L['block_mask'].q_indices, Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.int32, device=0, requires_grad=False, size=[1, 1, 16, 16], stride=[256, 256, 16, 1]) | | | +- NO_HASATTR: hasattr(L['block_mask'].q_indices, '_dynamo_dynamic_indices') == False | | | +- NO_TENSOR_ALIASING | | +- GuardManager: source=L['block_mask'].BLOCK_SIZE, accessed_by=GetAttrGuardAccessor(BLOCK_SIZE) - | | | +- TYPE_MATCH: ___check_type_id(L['block_mask'].BLOCK_SIZE, 8815232) + | | | +- TYPE_MATCH: ___check_type_id(L['block_mask'].BLOCK_SIZE, 8812224) | | | +- LENGTH_CHECK: len(L['block_mask'].BLOCK_SIZE) == 2 | | | +- GuardManager: source=L['block_mask'].BLOCK_SIZE[0], accessed_by=TupleGetItemGuardAccessor(0) | | | | +- EQUALS_MATCH: L['block_mask'].BLOCK_SIZE[0] == 128 @@ -3475,48 +4739,54 @@ V1003 10:11:03.320000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | +- GuardManager: source=L['block_mask'].as_tuple, accessed_by=GetAttrGuardAccessor(as_tuple) | | | +- GuardManager: source=L['block_mask'].as_tuple, accessed_by=FuncDefaultsGuardAccessor | | | | +- GuardManager: source=L['block_mask'].as_tuple.__defaults__[0], accessed_by=GetItemGuardAccessor(0) - | | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].as_tuple.__defaults__[0], 8911040) - | +- GuardManager: source=L['flex_attention'], accessed_by=DictGetItemGuardAccessor(flex_attention) + | | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].as_tuple.__defaults__[0], 8908032) + | +- GuardManager: source=L['flex_attention'], accessed_by=DictGetItemGuardAccessor('flex_attention') | | +- GuardManager: source=L['flex_attention'].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__code__, 387082992) + | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__code__, 139318784) | | +- GuardManager: source=L['flex_attention'], accessed_by=FuncDefaultsGuardAccessor | | | +- GuardManager: source=L['flex_attention'].__defaults__[2], accessed_by=GetItemGuardAccessor(2) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[2], 8825760) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[2], 8822752) | | | +- GuardManager: source=L['flex_attention'].__defaults__[3], accessed_by=GetItemGuardAccessor(3) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[3], 8910592) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[3], 8907584) | | | +- GuardManager: source=L['flex_attention'].__defaults__[4], accessed_by=GetItemGuardAccessor(4) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[4], 8910592) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[4], 8907584) | | | +- GuardManager: source=L['flex_attention'].__defaults__[5], accessed_by=GetItemGuardAccessor(5) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[5], 8825760) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[5], 8822752) | +- GuardManager: source=G, accessed_by=GlobalsGuardAccessor - | | +- GuardManager: source=G['__builtins_dict___6'], accessed_by=DictGetItemGuardAccessor(__builtins_dict___6) - | | | +- GuardManager: source=G['__builtins_dict___6']['len'], accessed_by=DictGetItemGuardAccessor(len) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['len'], 140413275558816) - | | | +- GuardManager: source=G['__builtins_dict___6']['sum'], accessed_by=DictGetItemGuardAccessor(sum) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['sum'], 140413275559936) - | | | +- GuardManager: source=G['__builtins_dict___6']['list'], accessed_by=DictGetItemGuardAccessor(list) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['list'], 8844320) - | | | +- GuardManager: source=G['__builtins_dict___6']['type'], accessed_by=DictGetItemGuardAccessor(type) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['type'], 8813248) - | | | +- GuardManager: source=G['__builtins_dict___6']['tuple'], accessed_by=DictGetItemGuardAccessor(tuple) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['tuple'], 8815232) - | | | +- GuardManager: source=G['__builtins_dict___6']['object'], accessed_by=DictGetItemGuardAccessor(object) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['object'], 8813984) - | | | +- GuardManager: source=G['__builtins_dict___6']['isinstance'], accessed_by=DictGetItemGuardAccessor(isinstance) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['isinstance'], 140413275558496) - | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_utils_dot__pytree) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140411217627952) + | | +- GuardManager: source=G['_140062263790704_c3'], accessed_by=DictGetItemGuardAccessor('_140062263790704_c3') + | | | +- GuardManager: source=G['_140062263790704_c3'].Tensor, accessed_by=GetAttrGuardAccessor(Tensor) + | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c3'].Tensor, 82181376) + | | | | +- GuardManager: source=G['_140062263790704_c3'].Tensor.__bases__, accessed_by=GetAttrGuardAccessor(__bases__) + | | | | | +- GuardManager: source=G['_140062263790704_c3'].Tensor.__bases__[0], accessed_by=TupleGetItemGuardAccessor(0) + | | | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c3'].Tensor.__bases__[0], 140062119703136) + | | +- GuardManager: source=G['__builtins_dict___6'], accessed_by=DictGetItemGuardAccessor('__builtins_dict___6') + | | | +- GuardManager: source=G['__builtins_dict___6']['len'], accessed_by=DictGetItemGuardAccessor('len') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['len'], 140062269592480) + | | | +- GuardManager: source=G['__builtins_dict___6']['sum'], accessed_by=DictGetItemGuardAccessor('sum') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['sum'], 140062269593600) + | | | +- GuardManager: source=G['__builtins_dict___6']['list'], accessed_by=DictGetItemGuardAccessor('list') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['list'], 8841312) + | | | +- GuardManager: source=G['__builtins_dict___6']['type'], accessed_by=DictGetItemGuardAccessor('type') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['type'], 8810240) + | | | +- GuardManager: source=G['__builtins_dict___6']['tuple'], accessed_by=DictGetItemGuardAccessor('tuple') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['tuple'], 8812224) + | | | +- GuardManager: source=G['__builtins_dict___6']['object'], accessed_by=DictGetItemGuardAccessor('object') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['object'], 8810976) + | | | +- GuardManager: source=G['__builtins_dict___6']['isinstance'], accessed_by=DictGetItemGuardAccessor('isinstance') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___6']['isinstance'], 140062269592160) + | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_utils_dot__pytree') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140057600329744) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].TreeSpec, accessed_by=GetAttrGuardAccessor(TreeSpec) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 84866496) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 86791872) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf, accessed_by=GetAttrGuardAccessor(_is_leaf) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140411217262720) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140057601749808) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, accessed_by=GetAttrGuardAccessor(_LEAF_SPEC) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 85171104) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 87025712) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, accessed_by=GetAttrGuardAccessor(type) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, accessed_by=GetAttrGuardAccessor(context) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes, accessed_by=GetAttrGuardAccessor(num_nodes) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes == 1 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_leaves, accessed_by=GetAttrGuardAccessor(num_leaves) @@ -3524,37 +4794,37 @@ V1003 10:11:03.320000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children, accessed_by=GetAttrGuardAccessor(num_children) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children == 0 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, accessed_by=GetAttrGuardAccessor(children_specs) - | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8844320) + | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8841312) | | | | | +- LENGTH_CHECK: not G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type, accessed_by=GetAttrGuardAccessor(_get_node_type) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140411217262448) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140057601749536) | | | +- DictGuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES, accessed_by=GetAttrGuardAccessor(SUPPORTED_NODES) - | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 519596 + | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 366913 | | | | +- KeyValueManager pair at index=1 | | | | | +- ValueManager: GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]] | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn, accessed_by=GetAttrGuardAccessor(flatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140411196281984) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140057599989696) | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn, accessed_by=GetAttrGuardAccessor(unflatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140411217182288) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140057600267792) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper, accessed_by=GetAttrGuardAccessor(_tree_flatten_helper) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140411217413040) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140057600001392) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance, accessed_by=GetAttrGuardAccessor(_is_namedtuple_instance) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140411217412592) - | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_comptime'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot__dynamo_dot_comptime) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_comptime'], 140410226912176) - | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_decorators'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot__dynamo_dot_decorators) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'], 140410226910096) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140057600000944) + | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_comptime'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot__dynamo_dot_comptime') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_comptime'], 140057359527872) + | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_decorators'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot__dynamo_dot_decorators') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'], 140057359526192) | | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_decorators'].is_compiling, accessed_by=GetAttrGuardAccessor(is_compiling) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'].is_compiling, 140410376252096) - | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_nn_dot_attention_dot__utils) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot__utils'], 140409673896784) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'].is_compiling, 140057422814624) + | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_nn_dot_attention_dot__utils') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot__utils'], 140048551571072) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS, accessed_by=GetAttrGuardAccessor(_SUPPORTED_HEAD_DIMS) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS, 8844320) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS, 8841312) | | | | +- LENGTH_CHECK: len(G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS) == 10 | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[0], accessed_by=ListGetItemGuardAccessor(0) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[0] == 2 @@ -3568,131 +4838,155 @@ V1003 10:11:03.320000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[4] == 32 | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[5], accessed_by=ListGetItemGuardAccessor(5) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[5] == 64 - | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_nn_dot_attention_dot_flex_attention) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], 140409673895824) + | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_nn_dot_attention_dot_flex_attention') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], 140048551568912) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math, accessed_by=GetAttrGuardAccessor(math) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math, 140413266939392) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math, 140062267404384) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math.sqrt, accessed_by=GetAttrGuardAccessor(sqrt) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math.sqrt, 140413266943072) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math.sqrt, 140062267408144) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch, accessed_by=GetAttrGuardAccessor(torch) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch, 140413267918368) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch, 140062263790704) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo, accessed_by=GetAttrGuardAccessor(_dynamo) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo, 140413260098400) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo, 140062262057760) | | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static, accessed_by=GetAttrGuardAccessor(mark_static) | | | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static.__code__, 123166432) + | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static.__code__, 125680000) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler, accessed_by=GetAttrGuardAccessor(compiler) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler, 140410826010400) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler, 140057571877776) | | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler.is_dynamo_compiling, accessed_by=GetAttrGuardAccessor(is_dynamo_compiling) - | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler.is_dynamo_compiling, 140410826132992) + | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler.is_dynamo_compiling, 140057572232544) + | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.is_grad_enabled, accessed_by=GetAttrGuardAccessor(is_grad_enabled) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.is_grad_enabled, 140062253115936) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device, accessed_by=GetAttrGuardAccessor(_validate_device) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device.__code__, 140409673611088) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device.__code__, 140048552018544) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop, accessed_by=GetAttrGuardAccessor(flex_attention_hop) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop, 96992544) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop, 96230624) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop.__name__, accessed_by=GetAttrGuardAccessor(__name__) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop.__name__ == 'flex_attention' | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim, accessed_by=GetAttrGuardAccessor(_supported_head_dim) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim.__code__, 140409673231376) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim.__code__, 140048552864992) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim, accessed_by=GetAttrGuardAccessor(_validate_embed_dim) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim.__code__, 388086512) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim.__code__, 139781872) + | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_nestedness, accessed_by=GetAttrGuardAccessor(_validate_nestedness) + | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_nestedness.__code__, accessed_by=GetAttrGuardAccessor(__code__) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_nestedness.__code__, 140048553100560) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input, accessed_by=GetAttrGuardAccessor(_validate_sdpa_input) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input.__code__, 387915104) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input.__code__, 139823744) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options, accessed_by=GetAttrGuardAccessor(_apply_kernel_options) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options.__code__, 140409683680752) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options.__code__, 140048554555888) -V1003 10:11:03.321000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "341d7b7d938fce7537065544860a0efd"} +V1206 15:24:50.619000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b414209b816a551182c3cbb3ebebb2ce"} { "name": "entire_frame_compile", - "ts": 1727975463321242.8, + "ts": 1733527490619189.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "_compile.compile_inner", + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.321000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "02dcbee7c15ab2085e7ee42a1679beb8"} +V1206 15:24:50.621000 1667746 torch/_dynamo/utils.py:1045] {"compilation_metrics": {"pre_grad_pass_time_us": 521, "joint_graph_pass_time_us": 1048, "feature_usage": {"pytorch/remote_cache:fx_graph_memcache_version": true}, "distributed_ephemeral_timeout_us": 4009141, "inductor_cumulative_compile_time_us": 77456, "aot_autograd_cumulative_compile_time_us": 134960, "dynamo_cumulative_compile_time_us": 242984, "frame_key": "2", "co_name": "fn", "co_filename": "/data/users/xmfan/a/pytorch/test/inductor/test_codecache.py", "co_firstlineno": 459, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 91, "shape_env_guard_count": 0, "graph_op_count": 2, "graph_node_count": 16, "graph_input_count": 11, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": false, "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}", "is_forward": true, "dynamo_compile_time_before_restart_us": 0}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.622000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c94fa0f0ebcc0c15f083237f9bd40372"} { - "name": "_compile.compile_inner", - "ts": 1727975463321536.0, + "name": "dynamo", + "ts": 1733527490621989.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0", + "frame_key": "2", + "co_name": "fn", + "co_filename": "/data/users/xmfan/a/pytorch/test/inductor/test_codecache.py", + "co_firstlineno": 459, + "cache_size": 0, + "accumulated_cache_size": 0, + "guard_count": 91, + "shape_env_guard_count": 0, + "graph_op_count": 2, + "graph_node_count": 16, + "graph_input_count": 11, + "fail_type": null, + "fail_reason": null, + "fail_user_frame_filename": null, + "fail_user_frame_lineno": null, + "non_compliant_ops": [], + "compliant_custom_ops": [], + "restart_reasons": [], + "dynamo_time_before_restart_s": 0.0, + "has_guarded_code": true, + "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.321000 2235078 torch/_dynamo/utils.py:840] {"compilation_metrics": {"compile_id": "1/0", "frame_key": "2", "co_name": "fn", "co_filename": "/data/users/oulgen/pytorch/test/inductor/test_codecache.py", "co_firstlineno": 379, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 80, "shape_env_guard_count": 0, "graph_op_count": 11, "graph_node_count": 25, "graph_input_count": 11, "start_time": 1727975462.8240004, "entire_frame_compile_time_s": 0.4971275329589844, "backend_compile_time_s": 0.2832028865814209, "inductor_compile_time_s": 0.1434309482574463, "code_gen_time_s": null, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "possibly_missed_reinplacing_opportunities": 0, "remote_cache_time_saved_s": 6.372999079, "structured_logging_overhead_s": 0.055921114, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": true}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.325000 2235078 torch/_dynamo/convert_frame.py:915] {"dynamo_start": {"stack": [{"line": 916, "name": "", "filename": 1}, {"line": 14, "name": "run_tests", "filename": 2}, {"line": 38, "name": "run_tests", "filename": 3}, {"line": 1273, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3116, "name": "run", "filename": 4}, {"line": 3088, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 2983, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 414, "name": "test_flex_attention_caching", "filename": 1}, {"line": 386, "name": "fn2", "filename": 1}, {"line": 1062, "name": "flex_attention", "filename": 10}, {"line": 1049, "name": "_flex_attention_hop_wrapper", "filename": 10}]}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.325000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "af0d73822951ce32af4202533e2a8290"} +V1206 15:24:50.623000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "f6ab5cb1835b6c744bb2f21749f1693d"} { - "name": "_compile.compile_inner", - "ts": 1727975463325378.5, - "args": null, + "name": "dynamo", + "ts": 1733527490623683.8, + "args": { + "compile_id": "0/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.325000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "dc5cd9b8a50a9005377afd5d28ba36fb"} +V1206 15:24:50.624000 1667746 torch/_dynamo/convert_frame.py:961] {"dynamo_start": {"stack": [{"line": 1439, "name": "", "filename": 1}, {"line": 15, "name": "run_tests", "filename": 2}, {"line": 39, "name": "run_tests", "filename": 3}, {"line": 1353, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3234, "name": "run", "filename": 4}, {"line": 3206, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 3099, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 490, "name": "test_flex_attention_caching", "filename": 1}, {"line": 466, "name": "fn2", "filename": 1}, {"line": 1316, "name": "flex_attention", "filename": 10}, {"line": 1303, "name": "_flex_attention_hop_wrapper", "filename": 10}]}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.624000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "0a08617336a65119a060bbb5cc88acf6"} { "name": "entire_frame_compile", - "ts": 1727975463325378.5, - "args": null, + "ts": 1733527490624241.5, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.328000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 348, "size": 524288}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.328000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.328000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 0, "source": "L['args'][0]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.441000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 1, "describer_id": 348, "size": 524288}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.442000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 2, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.442000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 2, "source": "L['args'][1]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.443000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 2, "describer_id": 348, "size": 524288}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.444000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 3, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.444000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 3, "source": "L['args'][2]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.445000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 3, "describer_id": 348, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.445000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 4, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.445000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 4, "source": "L['args'][4][0]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.446000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 4, "describer_id": 348, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.447000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 5, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.447000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 5, "source": "L['args'][4][1]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.448000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 5, "describer_id": 348, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.448000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 6, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.448000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 6, "source": "L['args'][4][2]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.449000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 6, "describer_id": 348, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.450000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.450000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 7, "source": "L['args'][4][3]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.451000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 7, "describer_id": 348, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.451000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.451000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 8, "source": "L['args'][4][4]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.452000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 8, "describer_id": 348, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.453000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.453000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 9, "source": "L['args'][4][5]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.454000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 9, "describer_id": 348, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.454000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.455000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 10, "source": "L['args'][4][6]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.456000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 10, "describer_id": 348, "size": 1024}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.456000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 348}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.456000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 348, "id": 11, "source": "L['args'][4][7]"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.464000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_output_graph": {"sizes": {"l_args_0_": [1, 4, 512, 64], "l_args_1_": [1, 4, 512, 64], "l_args_2_": [1, 4, 512, 64], "l_args_4_0_": [1, 1, 16], "l_args_4_1_": [1, 1, 16, 16], "l_args_4_2_": [1, 1, 16], "l_args_4_3_": [1, 1, 16, 16], "l_args_4_4_": [1, 1, 16], "l_args_4_5_": [1, 1, 16, 16], "l_args_4_6_": [1, 1, 16], "l_args_4_7_": [1, 1, 16, 16], "child_1": [], "child_2": [], "child_3": [], "child_4": [], "child": [], "child_5": [], "child_6": [], "child_7": [], "child_8": [], "getitem": [1, 4, 512, 64], "getitem_1": [1, 4, 512]}}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "c759d677da3d86354d83c3018e999089"} +V1206 15:24:50.625000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 0, "describer_id": 320, "size": 524288}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.626000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.626000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 0, "source": "L['args'][0]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.686000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 1, "describer_id": 320, "size": 524288}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.686000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 2, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.686000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 2, "source": "L['args'][1]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.687000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 2, "describer_id": 320, "size": 524288}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.687000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 3, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.687000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 3, "source": "L['args'][2]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.688000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 3, "describer_id": 320, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.688000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 4, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.688000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 4, "source": "L['args'][4][0]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.688000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 4, "describer_id": 320, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.688000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 5, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.689000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 5, "source": "L['args'][4][1]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.689000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 5, "describer_id": 320, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.689000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 6, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.689000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 6, "source": "L['args'][4][2]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.690000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 6, "describer_id": 320, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.690000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.690000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 7, "source": "L['args'][4][3]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.691000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 7, "describer_id": 320, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.691000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.691000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 8, "source": "L['args'][4][4]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.691000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 8, "describer_id": 320, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.692000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.692000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 9, "source": "L['args'][4][5]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.692000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 9, "describer_id": 320, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.692000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.692000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 10, "source": "L['args'][4][6]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.693000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 10, "describer_id": 320, "size": 1024}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.693000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 320}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.693000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 320, "id": 11, "source": "L['args'][4][7]"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.697000 1667746 torch/_dynamo/output_graph.py:1336] {"dynamo_output_graph": {"sizes": {"l_args_0_": [1, 4, 512, 64], "l_args_1_": [1, 4, 512, 64], "l_args_2_": [1, 4, 512, 64], "l_args_4_0_": [1, 1, 16], "l_args_4_1_": [1, 1, 16, 16], "l_args_4_2_": [1, 1, 16], "l_args_4_3_": [1, 1, 16, 16], "l_args_4_4_": [1, 1, 16], "l_args_4_5_": [1, 1, 16, 16], "l_args_4_6_": [1, 1, 16], "l_args_4_7_": [1, 1, 16, 16], "getitem": [1, 4, 512, 64], "getitem_1": [1, 4, 512]}}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "9bb6d56ee069d3045a9d8d21bcfdcd31"} class GraphModule(torch.nn.Module): def forward(self, L_args_0_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_args_1_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_args_2_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_args_4_0_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_1_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_args_4_2_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_3_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_args_4_4_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_5_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_args_4_6_: "i32[1, 1, 16][16, 16, 1]cuda:0", L_args_4_7_: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): l_args_0_ = L_args_0_ @@ -3707,19 +5001,10 @@ V1003 10:11:03.464000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_outpu l_args_4_6_ = L_args_4_6_ l_args_4_7_ = L_args_4_7_ - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1050 in _flex_attention_hop_wrapper, code: return flex_attention_hop(*args, **kwargs) - child_1: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_1 = None - child_2: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_2 = None - child_3: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_3 = None - child_4: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_4 = None - child: "f32[][]cuda:0" = l_args_0_.new_empty([], requires_grad = False); child = None + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1304 in _flex_attention_hop_wrapper, code: return flex_attention_hop(*args, **kwargs) score_mod_0 = self.score_mod_0 - child_5: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_5 = None - child_6: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_6 = None - child_7: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_7 = None - child_8: "i32[][]cuda:0" = l_args_0_.new_empty([], dtype = torch.int32); child_8 = None mask_fn_0 = self.mask_fn_0 - flex_attention = torch.ops.higher_order.flex_attention(l_args_0_, l_args_1_, l_args_2_, score_mod_0, (l_args_4_0_, l_args_4_1_, l_args_4_2_, l_args_4_3_, l_args_4_4_, l_args_4_5_, l_args_4_6_, l_args_4_7_, 128, 128, mask_fn_0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); l_args_0_ = l_args_1_ = l_args_2_ = score_mod_0 = l_args_4_0_ = l_args_4_1_ = l_args_4_2_ = l_args_4_3_ = l_args_4_4_ = l_args_4_5_ = l_args_4_6_ = l_args_4_7_ = mask_fn_0 = None + flex_attention = torch.ops.higher_order.flex_attention(l_args_0_, l_args_1_, l_args_2_, score_mod_0, (l_args_4_0_, l_args_4_1_, l_args_4_2_, l_args_4_3_, l_args_4_4_, l_args_4_5_, l_args_4_6_, l_args_4_7_, 128, 128, mask_fn_0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); l_args_0_ = l_args_1_ = l_args_2_ = score_mod_0 = l_args_4_0_ = l_args_4_1_ = l_args_4_2_ = l_args_4_3_ = l_args_4_4_ = l_args_4_5_ = l_args_4_6_ = l_args_4_7_ = mask_fn_0 = None getitem: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0] getitem_1: "f32[1, 4, 512][2048, 512, 1]cuda:0" = flex_attention[1]; flex_attention = None return (getitem, getitem_1) @@ -3729,72 +5014,46 @@ V1003 10:11:03.464000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_outpu return child class mask_fn_0(torch.nn.Module): - def forward(self, child_5: "i32[][]cuda:0", child_6: "i32[][]cuda:0", child_7: "i32[][]cuda:0", child_8: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 - ge: "b8[][]cuda:0" = child_7 >= child_8; child_7 = child_8 = None + def forward(self, child: "i32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0"): + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + ge: "b8[][]cuda:0" = child_2 >= child_3; child_2 = child_3 = None return ge -V1003 10:11:03.465000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "3f2a0e3577c204c69d66ef7550317ca5"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975463465456.5, - "args": null, - "ph": "B", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0 - } -V1003 10:11:03.465000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "b48e24df1eadf9ac86a2058fefc95dfb"} - { - "name": "backend_compile", - "ts": 1727975463465456.5, - "args": null, - "ph": "B", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0 - } -V1003 10:11:03.466000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "9ed8f1c18e744cba5c7e86e34c900750"} +V1206 15:24:50.698000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "c8302b73abb12ac1bce37cef0c47d244"} { "name": "backend_compile", - "ts": 1727975463466215.0, + "ts": 1733527490698251.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" }, - "ph": "E", + "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.466000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "c19895e106748f6b125e2df40adb8b9e"} +V1206 15:24:50.698000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "6b9dbcf1651e5e99de517f92d22baa7b"} { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975463466480.0, + "name": "backend_compile", + "ts": 1733527490698529.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.490000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_str": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "6c94a4177e3b74bb7d4cee6cdd33898c"} +V1206 15:24:50.709000 1667746 torch/_dynamo/guards.py:2315] {"dynamo_cpp_guards_str": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "a32f9935e2d4f1dc06e2de195d9c24b2"} TREE_GUARD_MANAGER: +- RootGuardManager - | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:471 in init_ambient_guards + | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:484 in init_ambient_guards | +- GLOBAL_STATE: ___check_global_state() | +- TORCH_FUNCTION_MODE_STACK: ___check_torch_function_mode_stack() - | +- GuardManager: source=L['args'], accessed_by=DictGetItemGuardAccessor(args) - | | +- TYPE_MATCH: ___check_type_id(L['args'], 8815232) + | +- GuardManager: source=L['args'], accessed_by=DictGetItemGuardAccessor('args') + | | +- TYPE_MATCH: ___check_type_id(L['args'], 8812224) | | +- LENGTH_CHECK: len(L['args']) == 7 | | +- GuardManager: source=L['args'][0], accessed_by=TupleGetItemGuardAccessor(0) | | | +- TENSOR_MATCH: check_tensor(L['args'][0], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) @@ -3810,9 +5069,9 @@ V1003 10:11:03.490000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | +- NO_TENSOR_ALIASING | | +- GuardManager: source=L['args'][3], accessed_by=TupleGetItemGuardAccessor(3) | | | +- GuardManager: source=L['args'][3].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][3].__code__, 140413275216112) + | | | | +- ID_MATCH: ___check_obj_id(L['args'][3].__code__, 140062269255152) | | +- GuardManager: source=L['args'][4], accessed_by=TupleGetItemGuardAccessor(4) - | | | +- TYPE_MATCH: ___check_type_id(L['args'][4], 8815232) + | | | +- TYPE_MATCH: ___check_type_id(L['args'][4], 8812224) | | | +- LENGTH_CHECK: len(L['args'][4]) == 11 | | | +- GuardManager: source=L['args'][4][0], accessed_by=TupleGetItemGuardAccessor(0) | | | | +- TENSOR_MATCH: check_tensor(L['args'][4][0], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.int32, device=0, requires_grad=False, size=[1, 1, 16], stride=[16, 16, 1]) @@ -3852,52 +5111,60 @@ V1003 10:11:03.490000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | +- EQUALS_MATCH: L['args'][4][9] == 128 | | | +- GuardManager: source=L['args'][4][10], accessed_by=TupleGetItemGuardAccessor(10) | | | | +- GuardManager: source=L['args'][4][10].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(L['args'][4][10].__code__, 140413271880128) + | | | | | +- ID_MATCH: ___check_obj_id(L['args'][4][10].__code__, 140062268556768) | | +- GuardManager: source=L['args'][5], accessed_by=TupleGetItemGuardAccessor(5) | | | +- EQUALS_MATCH: L['args'][5] == 0.125 | | +- GuardManager: source=L['args'][6], accessed_by=TupleGetItemGuardAccessor(6) - | | | +- DICT_LENGTH: len(L['args'][6]) == 3 - | | | +- GuardManager: source=L['args'][6]['ROWS_GUARANTEED_SAFE'], accessed_by=DictGetItemGuardAccessor(ROWS_GUARANTEED_SAFE) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['ROWS_GUARANTEED_SAFE'], 8910592) - | | | +- GuardManager: source=L['args'][6]['PRESCALE_QK'], accessed_by=DictGetItemGuardAccessor(PRESCALE_QK) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['PRESCALE_QK'], 8910592) - | | | +- GuardManager: source=L['args'][6]['OUTPUT_LOGSUMEXP'], accessed_by=DictGetItemGuardAccessor(OUTPUT_LOGSUMEXP) - | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['OUTPUT_LOGSUMEXP'], 8910592) - | +- GuardManager: source=L['kwargs'], accessed_by=DictGetItemGuardAccessor(kwargs) + | | | +- DICT_LENGTH: len(L['args'][6]) == 4 + | | | +- GuardManager: source=L['args'][6]['PRESCALE_QK'], accessed_by=DictGetItemGuardAccessor('PRESCALE_QK') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['PRESCALE_QK'], 8907584) + | | | +- GuardManager: source=L['args'][6]['ROWS_GUARANTEED_SAFE'], accessed_by=DictGetItemGuardAccessor('ROWS_GUARANTEED_SAFE') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['ROWS_GUARANTEED_SAFE'], 8907584) + | | | +- GuardManager: source=L['args'][6]['BLOCKS_ARE_CONTIGUOUS'], accessed_by=DictGetItemGuardAccessor('BLOCKS_ARE_CONTIGUOUS') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['BLOCKS_ARE_CONTIGUOUS'], 8907584) + | | | +- GuardManager: source=L['args'][6]['OUTPUT_LOGSUMEXP'], accessed_by=DictGetItemGuardAccessor('OUTPUT_LOGSUMEXP') + | | | | +- ID_MATCH: ___check_obj_id(L['args'][6]['OUTPUT_LOGSUMEXP'], 8908032) + | +- GuardManager: source=L['kwargs'], accessed_by=DictGetItemGuardAccessor('kwargs') | | +- DICT_LENGTH: not L['kwargs'] | +- GuardManager: source=G, accessed_by=GlobalsGuardAccessor - | | +- GuardManager: source=G['flex_attention_hop'], accessed_by=DictGetItemGuardAccessor(flex_attention_hop) - | | | +- TYPE_MATCH: ___check_type_id(G['flex_attention_hop'], 96992544) + | | +- GuardManager: source=G['flex_attention_hop'], accessed_by=DictGetItemGuardAccessor('flex_attention_hop') + | | | +- TYPE_MATCH: ___check_type_id(G['flex_attention_hop'], 96230624) | | | +- GuardManager: source=G['flex_attention_hop'].__name__, accessed_by=GetAttrGuardAccessor(__name__) | | | | +- EQUALS_MATCH: G['flex_attention_hop'].__name__ == 'flex_attention' - | | +- GuardManager: source=G['__builtins_dict___8'], accessed_by=DictGetItemGuardAccessor(__builtins_dict___8) - | | | +- GuardManager: source=G['__builtins_dict___8']['len'], accessed_by=DictGetItemGuardAccessor(len) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['len'], 140413275558816) - | | | +- GuardManager: source=G['__builtins_dict___8']['sum'], accessed_by=DictGetItemGuardAccessor(sum) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['sum'], 140413275559936) - | | | +- GuardManager: source=G['__builtins_dict___8']['list'], accessed_by=DictGetItemGuardAccessor(list) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['list'], 8844320) - | | | +- GuardManager: source=G['__builtins_dict___8']['type'], accessed_by=DictGetItemGuardAccessor(type) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['type'], 8813248) - | | | +- GuardManager: source=G['__builtins_dict___8']['tuple'], accessed_by=DictGetItemGuardAccessor(tuple) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['tuple'], 8815232) - | | | +- GuardManager: source=G['__builtins_dict___8']['object'], accessed_by=DictGetItemGuardAccessor(object) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['object'], 8813984) - | | | +- GuardManager: source=G['__builtins_dict___8']['isinstance'], accessed_by=DictGetItemGuardAccessor(isinstance) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['isinstance'], 140413275558496) - | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_utils_dot__pytree) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140411217627952) + | | +- GuardManager: source=G['_140062263790704_c4'], accessed_by=DictGetItemGuardAccessor('_140062263790704_c4') + | | | +- GuardManager: source=G['_140062263790704_c4'].Tensor, accessed_by=GetAttrGuardAccessor(Tensor) + | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c4'].Tensor, 82181376) + | | | | +- GuardManager: source=G['_140062263790704_c4'].Tensor.__bases__, accessed_by=GetAttrGuardAccessor(__bases__) + | | | | | +- GuardManager: source=G['_140062263790704_c4'].Tensor.__bases__[0], accessed_by=TupleGetItemGuardAccessor(0) + | | | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c4'].Tensor.__bases__[0], 140062119703136) + | | +- GuardManager: source=G['__builtins_dict___8'], accessed_by=DictGetItemGuardAccessor('__builtins_dict___8') + | | | +- GuardManager: source=G['__builtins_dict___8']['len'], accessed_by=DictGetItemGuardAccessor('len') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['len'], 140062269592480) + | | | +- GuardManager: source=G['__builtins_dict___8']['sum'], accessed_by=DictGetItemGuardAccessor('sum') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['sum'], 140062269593600) + | | | +- GuardManager: source=G['__builtins_dict___8']['list'], accessed_by=DictGetItemGuardAccessor('list') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['list'], 8841312) + | | | +- GuardManager: source=G['__builtins_dict___8']['type'], accessed_by=DictGetItemGuardAccessor('type') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['type'], 8810240) + | | | +- GuardManager: source=G['__builtins_dict___8']['tuple'], accessed_by=DictGetItemGuardAccessor('tuple') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['tuple'], 8812224) + | | | +- GuardManager: source=G['__builtins_dict___8']['object'], accessed_by=DictGetItemGuardAccessor('object') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['object'], 8810976) + | | | +- GuardManager: source=G['__builtins_dict___8']['isinstance'], accessed_by=DictGetItemGuardAccessor('isinstance') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___8']['isinstance'], 140062269592160) + | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_utils_dot__pytree') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140057600329744) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].TreeSpec, accessed_by=GetAttrGuardAccessor(TreeSpec) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 84866496) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 86791872) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf, accessed_by=GetAttrGuardAccessor(_is_leaf) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140411217262720) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140057601749808) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, accessed_by=GetAttrGuardAccessor(_LEAF_SPEC) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 85171104) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 87025712) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, accessed_by=GetAttrGuardAccessor(type) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, accessed_by=GetAttrGuardAccessor(context) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes, accessed_by=GetAttrGuardAccessor(num_nodes) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes == 1 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_leaves, accessed_by=GetAttrGuardAccessor(num_leaves) @@ -3905,116 +5172,135 @@ V1003 10:11:03.490000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children, accessed_by=GetAttrGuardAccessor(num_children) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children == 0 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, accessed_by=GetAttrGuardAccessor(children_specs) - | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8844320) + | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8841312) | | | | | +- LENGTH_CHECK: not G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type, accessed_by=GetAttrGuardAccessor(_get_node_type) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140411217262448) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140057601749536) | | | +- DictGuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES, accessed_by=GetAttrGuardAccessor(SUPPORTED_NODES) - | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 519596 + | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 366913 | | | | +- KeyValueManager pair at index=1 | | | | | +- ValueManager: GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]] | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn, accessed_by=GetAttrGuardAccessor(flatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140411196281984) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140057599989696) | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn, accessed_by=GetAttrGuardAccessor(unflatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140411217182288) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140057600267792) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper, accessed_by=GetAttrGuardAccessor(_tree_flatten_helper) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140411217413040) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140057600001392) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance, accessed_by=GetAttrGuardAccessor(_is_namedtuple_instance) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140411217412592) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140057600000944) -V1003 10:11:03.491000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "eda00b5e5981dad4abbde14a07a38b9c"} +V1206 15:24:50.710000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "6d01a47ed965240f30333b7963f37f52"} { "name": "entire_frame_compile", - "ts": 1727975463491399.2, + "ts": 1733527490710248.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.491000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "dc3c58a766d43af7e167ce88702b4e5c"} +V1206 15:24:50.712000 1667746 torch/_dynamo/utils.py:1045] {"compilation_metrics": {"aot_autograd_cumulative_compile_time_us": 278, "dynamo_cumulative_compile_time_us": 86006, "frame_key": "1", "co_name": "_flex_attention_hop_wrapper", "co_filename": "/data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py", "co_firstlineno": 1303, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 58, "shape_env_guard_count": 0, "graph_op_count": 3, "graph_node_count": 17, "graph_input_count": 11, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": false, "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 9223372036854775807, \"accumulated_cache_size_limit\": 9223372036854775807, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}", "is_forward": true, "dynamo_compile_time_before_restart_us": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.713000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "bf4cb9e110f5e8f2910d1d589e140933"} { - "name": "_compile.compile_inner", - "ts": 1727975463491701.2, + "name": "dynamo", + "ts": 1733527490713016.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 1, - "fxgraph_cache_bypass": 0 - } + "compile_id": "0/0", + "frame_key": "1", + "co_name": "_flex_attention_hop_wrapper", + "co_filename": "/data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py", + "co_firstlineno": 1303, + "cache_size": 0, + "accumulated_cache_size": 0, + "guard_count": 58, + "shape_env_guard_count": 0, + "graph_op_count": 3, + "graph_node_count": 17, + "graph_input_count": 11, + "fail_type": null, + "fail_reason": null, + "fail_user_frame_filename": null, + "fail_user_frame_lineno": null, + "non_compliant_ops": [], + "compliant_custom_ops": [], + "restart_reasons": [], + "dynamo_time_before_restart_s": 0.0, + "has_guarded_code": true, + "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 9223372036854775807, \"accumulated_cache_size_limit\": 9223372036854775807, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.492000 2235078 torch/_dynamo/utils.py:840] {"compilation_metrics": {"compile_id": "0/0", "frame_key": "1", "co_name": "_flex_attention_hop_wrapper", "co_filename": "/data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py", "co_firstlineno": 1049, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 55, "shape_env_guard_count": 0, "graph_op_count": 12, "graph_node_count": 26, "graph_input_count": 11, "start_time": 1727975463.3253677, "entire_frame_compile_time_s": 0.16592144966125488, "backend_compile_time_s": 0.0006587505340576172, "inductor_compile_time_s": null, "code_gen_time_s": null, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "possibly_missed_reinplacing_opportunities": 0, "remote_cache_time_saved_s": 0, "structured_logging_overhead_s": 0.04651781599999999, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": true}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.496000 2235078 torch/_dynamo/convert_frame.py:915] {"dynamo_start": {"stack": [{"line": 916, "name": "", "filename": 1}, {"line": 14, "name": "run_tests", "filename": 2}, {"line": 38, "name": "run_tests", "filename": 3}, {"line": 1273, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3116, "name": "run", "filename": 4}, {"line": 3088, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 2983, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 414, "name": "test_flex_attention_caching", "filename": 1}, {"line": 385, "name": "fn2", "filename": 1}]}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.496000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "68d9557f97487c1ecd28ffaf4020aff9"} +V1206 15:24:50.715000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "34b5378127dd1adc3ac18035ff2352ab"} { - "name": "_compile.compile_inner", - "ts": 1727975463496392.5, - "args": null, + "name": "dynamo", + "ts": 1733527490715020.2, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.496000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "eab0483a7d9a0ee867d4596b4367f9e6"} +V1206 15:24:50.715000 1667746 torch/_dynamo/convert_frame.py:961] {"dynamo_start": {"stack": [{"line": 1439, "name": "", "filename": 1}, {"line": 15, "name": "run_tests", "filename": 2}, {"line": 39, "name": "run_tests", "filename": 3}, {"line": 1353, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3234, "name": "run", "filename": 4}, {"line": 3206, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 3099, "name": "wrapper", "filename": 4}, {"line": 81, "name": "inner", "filename": 9}, {"line": 81, "name": "inner", "filename": 9}, {"line": 490, "name": "test_flex_attention_caching", "filename": 1}, {"line": 465, "name": "fn2", "filename": 1}]}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.715000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "514fa302b9c8e853a86097889a21966b"} { "name": "entire_frame_compile", - "ts": 1727975463496392.5, - "args": null, + "ts": 1733527490715568.8, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.499000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 362, "size": 524288}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.499000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.499000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 0, "source": "L['q']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.508000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 1, "describer_id": 362, "size": 524288}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.508000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 1, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.509000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 1, "source": "L['k']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.510000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 2, "describer_id": 362, "size": 524288}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.510000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 2, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.511000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 2, "source": "L['v']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.520000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 3, "describer_id": 362, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.520000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 3, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.521000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 3, "source": "L['block_mask'].kv_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.653000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 4, "describer_id": 362, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.654000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 5, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.654000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 5, "source": "L['block_mask'].kv_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.655000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 5, "describer_id": 362, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.656000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 6, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.656000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 6, "source": "L['block_mask'].full_kv_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.657000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 6, "describer_id": 362, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.658000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.658000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 7, "source": "L['block_mask'].full_kv_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.659000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 7, "describer_id": 362, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.659000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.660000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 8, "source": "L['block_mask'].q_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.661000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 8, "describer_id": 362, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.661000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.661000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 9, "source": "L['block_mask'].q_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.663000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 9, "describer_id": 362, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.663000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.663000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 10, "source": "L['block_mask'].full_q_num_blocks"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.665000 2235078 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 10, "describer_id": 362, "size": 1024}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.665000 2235078 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 362}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.665000 2235078 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 362, "id": 11, "source": "L['block_mask'].full_q_indices"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V1003 10:11:03.673000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_output_graph": {"sizes": {"l_q_": [1, 4, 512, 64], "l_k_": [1, 4, 512, 64], "l_v_": [1, 4, 512, 64], "l_block_mask_kv_num_blocks": [1, 1, 16], "l_block_mask_kv_indices": [1, 1, 16, 16], "l_block_mask_full_kv_num_blocks": [1, 1, 16], "l_block_mask_full_kv_indices": [1, 1, 16, 16], "l_block_mask_q_num_blocks": [1, 1, 16], "l_block_mask_q_indices": [1, 1, 16, 16], "l_block_mask_full_q_num_blocks": [1, 1, 16], "l_block_mask_full_q_indices": [1, 1, 16, 16], "child_1": [], "child_2": [], "child_3": [], "child_4": [], "child": [], "child_5": [], "child_6": [], "child_7": [], "child_8": [], "out": [1, 4, 512, 64]}}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "487c69baa1e2981125efdf989466d11f"} +V1206 15:24:50.716000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 0, "describer_id": 335, "size": 524288}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.717000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 0, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.717000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 0, "source": "L['q']"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.723000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 1, "describer_id": 335, "size": 524288}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.723000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 1, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 1, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.723000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 1, "source": "L['k']"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.724000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 2, "describer_id": 335, "size": 524288}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.724000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 2, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 4, 512, 64], "is_leaf": true, "stride": [131072, 32768, 64, 1], "storage": 2, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.724000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 2, "source": "L['v']"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.729000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 3, "describer_id": 335, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.730000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 3, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 3, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.730000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 3, "source": "L['block_mask'].kv_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.795000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 4, "describer_id": 335, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.795000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 5, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 4, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.796000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 5, "source": "L['block_mask'].kv_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.796000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 5, "describer_id": 335, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.797000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 6, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 5, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.797000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 6, "source": "L['block_mask'].full_kv_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.798000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 6, "describer_id": 335, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.798000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 7, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 6, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.798000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 7, "source": "L['block_mask'].full_kv_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.799000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 7, "describer_id": 335, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.799000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 8, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 7, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.799000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 8, "source": "L['block_mask'].q_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.800000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 8, "describer_id": 335, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.800000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 9, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 8, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.800000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 9, "source": "L['block_mask'].q_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.801000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 9, "describer_id": 335, "size": 64}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.801000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 10, "ndim": 3, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16], "is_leaf": true, "stride": [16, 16, 1], "storage": 9, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.801000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 10, "source": "L['block_mask'].full_q_num_blocks"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.802000 1667746 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 10, "describer_id": 335, "size": 1024}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.802000 1667746 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 11, "ndim": 4, "dtype": "torch.int32", "device": "device(type='cuda', index=0)", "size": [1, 1, 16, 16], "is_leaf": true, "stride": [256, 256, 16, 1], "storage": 10, "view_func": "", "describer_id": 335}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.802000 1667746 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 335, "id": 11, "source": "L['block_mask'].full_q_indices"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:50.806000 1667746 torch/_dynamo/output_graph.py:1336] {"dynamo_output_graph": {"sizes": {"l_q_": [1, 4, 512, 64], "l_k_": [1, 4, 512, 64], "l_v_": [1, 4, 512, 64], "l_block_mask_kv_num_blocks": [1, 1, 16], "l_block_mask_kv_indices": [1, 1, 16, 16], "l_block_mask_full_kv_num_blocks": [1, 1, 16], "l_block_mask_full_kv_indices": [1, 1, 16, 16], "l_block_mask_q_num_blocks": [1, 1, 16], "l_block_mask_q_indices": [1, 1, 16, 16], "l_block_mask_full_q_num_blocks": [1, 1, 16], "l_block_mask_full_q_indices": [1, 1, 16, 16], "out": [1, 4, 512, 64]}}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "482a8ec46bfbc33e69d27b2c7e75e43d"} class GraphModule(torch.nn.Module): def forward(self, L_q_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_k_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_v_: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", L_block_mask_kv_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_kv_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_block_mask_full_kv_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_full_kv_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_block_mask_q_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_q_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", L_block_mask_full_q_num_blocks: "i32[1, 1, 16][16, 16, 1]cuda:0", L_block_mask_full_q_indices: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): l_q_ = L_q_ @@ -4029,19 +5315,10 @@ V1003 10:11:03.673000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_outpu l_block_mask_full_q_num_blocks = L_block_mask_full_q_num_blocks l_block_mask_full_q_indices = L_block_mask_full_q_indices - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( - child_1: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_1 = None - child_2: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_2 = None - child_3: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_3 = None - child_4: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_4 = None - child: "f32[][]cuda:0" = l_q_.new_empty([], requires_grad = False); child = None + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( score_mod_0 = self.score_mod_0 - child_5: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_5 = None - child_6: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_6 = None - child_7: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_7 = None - child_8: "i32[][]cuda:0" = l_q_.new_empty([], dtype = torch.int32); child_8 = None mask_fn_0 = self.mask_fn_0 - flex_attention = torch.ops.higher_order.flex_attention(l_q_, l_k_, l_v_, score_mod_0, (l_block_mask_kv_num_blocks, l_block_mask_kv_indices, l_block_mask_full_kv_num_blocks, l_block_mask_full_kv_indices, l_block_mask_q_num_blocks, l_block_mask_q_indices, l_block_mask_full_q_num_blocks, l_block_mask_full_q_indices, 128, 128, mask_fn_0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); l_q_ = l_k_ = l_v_ = score_mod_0 = l_block_mask_kv_num_blocks = l_block_mask_kv_indices = l_block_mask_full_kv_num_blocks = l_block_mask_full_kv_indices = l_block_mask_q_num_blocks = l_block_mask_q_indices = l_block_mask_full_q_num_blocks = l_block_mask_full_q_indices = mask_fn_0 = None + flex_attention = torch.ops.higher_order.flex_attention(l_q_, l_k_, l_v_, score_mod_0, (l_block_mask_kv_num_blocks, l_block_mask_kv_indices, l_block_mask_full_kv_num_blocks, l_block_mask_full_kv_indices, l_block_mask_q_num_blocks, l_block_mask_q_indices, l_block_mask_full_q_num_blocks, l_block_mask_full_q_indices, 128, 128, mask_fn_0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); l_q_ = l_k_ = l_v_ = score_mod_0 = l_block_mask_kv_num_blocks = l_block_mask_kv_indices = l_block_mask_full_kv_num_blocks = l_block_mask_full_kv_indices = l_block_mask_q_num_blocks = l_block_mask_q_indices = l_block_mask_full_q_num_blocks = l_block_mask_full_q_indices = mask_fn_0 = None out: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0]; flex_attention = None return (out,) @@ -4050,48 +5327,211 @@ V1003 10:11:03.673000 2235078 torch/_dynamo/output_graph.py:1347] {"dynamo_outpu return child class mask_fn_0(torch.nn.Module): - def forward(self, child_5: "i32[][]cuda:0", child_6: "i32[][]cuda:0", child_7: "i32[][]cuda:0", child_8: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 - ge: "b8[][]cuda:0" = child_7 >= child_8; child_7 = child_8 = None + def forward(self, child: "i32[][]cuda:0", child_1: "i32[][]cuda:0", child_2: "i32[][]cuda:0", child_3: "i32[][]cuda:0"): + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + ge: "b8[][]cuda:0" = child_2 >= child_3; child_2 = child_3 = None return ge -V1003 10:11:03.674000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "94724b556ec4c03944e43538685317ee"} +V1206 15:24:50.807000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "1d40dde81d7850073106ae3cf2229e48"} { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975463674357.2, - "args": null, + "name": "backend_compile", + "ts": 1733527490807005.5, + "args": { + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.674000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "726b102bddec88824ed82f86c8c02d39"} +V1206 15:24:50.807000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ca5c77d2199fb0261e09e1800f877003"} { - "name": "backend_compile", - "ts": 1727975463674357.2, - "args": null, + "name": "_recursive_pre_grad_passes", + "ts": 1733527490807277.8, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.680000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c103e7f8badb6398fb5a927c8aeef126"} +V1206 15:24:50.807000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "a8bfb8ba2a3fefc2b30a7110871d7e3a"} + { + "name": "_recursive_pre_grad_passes", + "ts": 1733527490807804.0, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.810000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "3b3ffc9d594503c9a35ba96fe32b6bde"} { "name": "create_aot_dispatcher_function", - "ts": 1727975463680196.2, - "args": null, + "ts": 1733527490809984.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.798000 2235078 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:215] {"aot_forward_graph": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "fe8d53ae08cd360e1378a8a527d186f0"} +V1206 15:24:50.852000 1667746 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:214] {"artifact": {"name": "aot_forward_graph_fw_metadata", "encoding": "string"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d952b3071dd9ba606a04d644841d9aba"} + ViewAndMutationMeta(input_info=[InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True), + InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True)], + output_info=[OutputAliasInfo(output_type=, + raw_type=, + base_idx=None, + dynamic_dims=set(), + requires_grad=False, + functional_tensor=None)], + num_intermediate_bases=0, + keep_input_mutations=True, + traced_tangents=[], + subclass_inp_meta=[PlainTensorMeta(unwrapped_idx=0, + memory_format=None), + PlainTensorMeta(unwrapped_idx=1, + memory_format=None), + PlainTensorMeta(unwrapped_idx=2, + memory_format=None), + PlainTensorMeta(unwrapped_idx=3, + memory_format=None), + PlainTensorMeta(unwrapped_idx=4, + memory_format=None), + PlainTensorMeta(unwrapped_idx=5, + memory_format=None), + PlainTensorMeta(unwrapped_idx=6, + memory_format=None), + PlainTensorMeta(unwrapped_idx=7, + memory_format=None), + PlainTensorMeta(unwrapped_idx=8, + memory_format=None), + PlainTensorMeta(unwrapped_idx=9, + memory_format=None), + PlainTensorMeta(unwrapped_idx=10, + memory_format=None)], + subclass_fw_graph_out_meta=[PlainTensorMeta(unwrapped_idx=0, + memory_format=None)], + subclass_tangent_meta=[], + is_train=False, + traced_tangent_metas=None, + num_symints_saved_for_bw=None, + grad_enabled_mutation=None, + deterministic=None, + static_input_indices=[], + tokens={}, + indices_of_inputs_that_requires_grad_with_mutations_in_bw=[], + bw_donated_idxs=None, + num_backward_tokens=0) +V1206 15:24:50.853000 1667746 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:232] {"aot_inference_graph": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ecc276ab6d6e35dc44c5f2bbc091ace0"} class (torch.nn.Module): def forward(self, arg0_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg1_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg2_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg3_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg4_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg5_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg6_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg7_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg8_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg9_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg10_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( sdpa_score0 = self.sdpa_score0 sdpa_mask0 = self.sdpa_mask0 - flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None + flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None getitem: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0]; flex_attention = None return (getitem,) @@ -4101,41 +5541,60 @@ V1003 10:11:03.798000 2235078 torch/_functorch/_aot_autograd/dispatch_and_compil class sdpa_mask0(torch.nn.Module): def forward(self, arg0_1: "i32[][]cuda:0", arg1_1: "i32[][]cuda:0", arg2_1: "i32[][]cuda:0", arg3_1: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 ge: "b8[][]cuda:0" = torch.ops.aten.ge.Tensor(arg2_1, arg3_1); arg2_1 = arg3_1 = None return ge -V1003 10:11:03.799000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "19af874521da2f24c9f873fb757d7dc6"} +V1206 15:24:50.853000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "de5dcb5ec710567996f5f45241f82616"} { "name": "compile_fx..fw_compiler_base", - "ts": 1727975463799121.2, - "args": null, + "ts": 1733527490853953.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.801000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "bb6048e2282795589bd2400ba093dfc8"} +V1206 15:24:50.854000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f7192ebd762dae9c8cf2d7972d6bf48e"} { - "name": "compile_fx_inner", - "ts": 1727975463801024.5, - "args": null, + "name": "_recursive_joint_graph_passes", + "ts": 1733527490854257.8, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.801000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b7b2a66f7ec1d6b6227990c8cd3b1d21"} +V1206 15:24:50.855000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "bfb4a0e07ad2c3524d5a0a3c69449196"} + { + "name": "_recursive_joint_graph_passes", + "ts": 1733527490855286.0, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.855000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "8d3b095b4ae2ad23dfdd31e5633971b0"} { "name": "inductor_compile", - "ts": 1727975463801024.5, - "args": null, + "ts": 1733527490855525.2, + "args": { + "fn_name": "compile_fx_inner", + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.816000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {"name": "fx_graph_runnable", "encoding": "string"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c507028eda974d39e4ca6593c6320460"} +V1206 15:24:50.863000 1667746 torch/_inductor/compile_fx.py:835] {"artifact": {"name": "fx_graph_runnable", "encoding": "string"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "2685e2fe8ae1c10315e3657921ebac7a"} import torch from torch import tensor, device @@ -4148,11 +5607,40 @@ V1003 10:11:03.816000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {" import torch._inductor.config import torch._functorch.config import torch.fx.experimental._config + torch._dynamo.config.cache_size_limit = 8 + torch._dynamo.config.accumulated_cache_size_limit = 256 + torch._dynamo.config.traceable_tensor_subclasses = set() + torch._dynamo.config.suppress_errors = False + torch._dynamo.config.allowed_functions_module_string_ignorelist = {'torch._decomp', 'torch.testing', 'torch._prims', 'torch._refs', 'torch.distributions'} + torch._dynamo.config._ddp_optimization_mode = ['ddp_optimizer', 'python_reducer', 'python_reducer_without_compiled_forward', 'no_optimization'] + torch._dynamo.config.raise_on_ctx_manager_usage = True + torch._dynamo.config._save_config_ignore = {'repro_after', 'skipfiles_inline_module_allowlist', 'repro_level', 'constant_functions'} torch._dynamo.config.log_compilation_metrics = False + torch._dynamo.config.reorderable_logging_functions = set() + torch._dynamo.config._autograd_backward_strict_mode_banned_ops = ['stride', 'requires_grad', 'storage_offset', 'layout', 'data', 'is_coalesced', 'is_complex', 'is_conj', 'is_contiguous', 'is_cpu', 'is_cuda', 'is_distributed', 'is_floating_point', 'is_inference', 'is_ipu', 'is_leaf', 'is_maia', 'is_meta', 'is_mkldnn', 'is_mps', 'is_mtia', 'is_neg', 'is_nested', 'is_nonzero', 'is_pinned', 'is_quantized', 'is_same_size', 'is_set_to', 'is_shared', 'is_signed', 'is_sparse', 'is_sparse_csr', 'is_vulkan', 'is_xla', 'is_xpu'] + torch._dynamo.config.fake_tensor_cache_enabled = True torch._dynamo.config.fake_tensor_cache_crosscheck_enabled = True + torch._dynamo.config.compiled_autograd_kwargs_override = {} + torch._inductor.config.fx_graph_cache = True torch._inductor.config.fx_graph_remote_cache = False torch._inductor.config.autotune_local_cache = False torch._inductor.config.autotune_remote_cache = False + torch._inductor.config.bundled_autotune_remote_cache = False + torch._inductor.config.pre_grad_fusion_options = {} + torch._inductor.config.post_grad_fusion_options = {} + torch._inductor.config.fx_passes_numeric_check = {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True} + torch._inductor.config.reorder_for_compute_comm_overlap_passes = ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms'] + torch._inductor.config._fuse_ddp_communication_passes = ['fuse_ddp_with_concat_op', 'schedule_comm_wait'] + torch._inductor.config.aot_inductor.metadata = {} + torch._inductor.config.aot_inductor.presets = {} + torch._inductor.config.rocm.arch = [] + torch._inductor.config.rocm.ck_supported_arch = ['gfx90a', 'gfx940', 'gfx941', 'gfx942'] + torch._inductor.config._save_config_ignore = ['trace.upload_tar', 'joint_custom_pre_pass', 'joint_custom_post_pass', 'pre_grad_custom_pass'] + torch._inductor.config._cache_config_ignore_prefix = ['trace', 'cuda.cutlass_dir', 'worker_start_method', 'compile_threads', 'post_grad_custom_post_pass', 'post_grad_custom_pre_pass', 'always_complex_memory_overlap_TESTING_ONLY'] + torch._inductor.config.external_matmul = [] + torch._functorch.config.functionalize_rng_ops = False + torch._functorch.config.enable_autograd_cache = True + torch._functorch.config.fake_tensor_allow_unsafe_data_ptr_access = True torch._functorch.config.unlift_effect_tokens = True @@ -4161,20 +5649,21 @@ V1003 10:11:03.816000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {" - # torch version: 2.5.0a0+git7647c39 - # torch cuda version: 12.0 - # torch git version: 7647c398ff87daf70260854cf0a7f7993b3abc76 + + # torch version: 2.6.0a0+giteece9ec + # torch cuda version: 12.2 + # torch git version: eece9ecd62cae84bc2f915fc48cffe43e30256aa # CUDA Info: # nvcc: NVIDIA (R) Cuda compiler driver # Copyright (c) 2005-2023 NVIDIA Corporation - # Built on Fri_Jan__6_16:45:21_PST_2023 - # Cuda compilation tools, release 12.0, V12.0.140 - # Build cuda_12.0.r12.0/compiler.32267302_0 + # Built on Tue_Aug_15_22:02:13_PDT_2023 + # Cuda compilation tools, release 12.2, V12.2.140 + # Build cuda_12.2.r12.2/compiler.33191640_0 # GPU Hardware Info: - # NVIDIA PG509-210 : 8 + # NVIDIA H100 : 8 from torch.nn import * @@ -4189,7 +5678,7 @@ V1003 10:11:03.816000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {" def forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1): sdpa_score0 = self.sdpa_score0 sdpa_mask0 = self.sdpa_mask0 - flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None + flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None getitem = flex_attention[0]; flex_attention = None return (getitem,) @@ -4225,13 +5714,37 @@ V1003 10:11:03.816000 2235078 torch/_inductor/compile_fx.py:731] {"artifact": {" # To run it separately, do # mod, args = run_repro(mod, load_args, accuracy=False, command='get_args', save_dir=None, tracing_mode='real', check_str=None) # mod(*args) -V1003 10:11:03.827000 2235078 torch/_inductor/compile_fx.py:795] {"inductor_post_grad_graph": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "fe8d53ae08cd360e1378a8a527d186f0"} +V1206 15:24:50.866000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "dfa5a083adca30c7330158b04eac504c"} + { + "name": "_recursive_post_grad_passes", + "ts": 1733527490866428.5, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.867000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "92fea8d9f2c2af6ab5a7b7726a25e9c7"} + { + "name": "_recursive_post_grad_passes", + "ts": 1733527490867486.8, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.868000 1667746 torch/_inductor/compile_fx.py:898] {"inductor_post_grad_graph": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ecc276ab6d6e35dc44c5f2bbc091ace0"} class (torch.nn.Module): def forward(self, arg0_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg1_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg2_1: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0", arg3_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg4_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg5_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg6_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg7_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg8_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0", arg9_1: "i32[1, 1, 16][16, 16, 1]cuda:0", arg10_1: "i32[1, 1, 16, 16][256, 256, 16, 1]cuda:0"): - # File: /data/users/oulgen/pytorch/torch/nn/attention/flex_attention.py:1032 in flex_attention, code: out, lse = flex_attention_hop( + # File: /data/users/xmfan/a/pytorch/torch/nn/attention/flex_attention.py:1286 in flex_attention, code: out, lse = flex_attention_hop( sdpa_score0 = self.sdpa_score0 sdpa_mask0 = self.sdpa_mask0 - flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None + flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None getitem: "f32[1, 4, 512, 64][131072, 32768, 64, 1]cuda:0" = flex_attention[0]; flex_attention = None return (getitem,) @@ -4241,135 +5754,180 @@ V1003 10:11:03.827000 2235078 torch/_inductor/compile_fx.py:795] {"inductor_post class sdpa_mask0(torch.nn.Module): def forward(self, arg0_1: "i32[][]cuda:0", arg1_1: "i32[][]cuda:0", arg2_1: "i32[][]cuda:0", arg3_1: "i32[][]cuda:0"): - # File: /data/users/oulgen/pytorch/test/inductor/test_codecache.py:373 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 + # File: /data/users/xmfan/a/pytorch/test/inductor/test_codecache.py:453 in , code: lambda b, h, q, kv: q >= kv, None, None, 2048, 2048 ge: "b8[][]cuda:0" = torch.ops.aten.ge.Tensor(arg2_1, arg3_1); arg2_1 = arg3_1 = None return ge -V1003 10:11:03.828000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "e24475885265d7add113de685b5b4785"} +V1206 15:24:50.869000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "2c4e2b8ec85ec846c7bfb5687ddbf4eb"} { "name": "GraphLowering.run", - "ts": 1727975463828251.2, - "args": null, + "ts": 1733527490869198.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.863000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "df613d910e7ef57b6d235421c34afbec"} +V1206 15:24:50.889000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "4cf953068399b7703999c88a61b534a8"} { "name": "GraphLowering.run", - "ts": 1727975463863685.0, + "ts": 1733527490889235.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.864000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "5fd8784b67f1838294fd144c5df7385a"} +V1206 15:24:50.889000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "8512f5af880647d00bdcc6cb55c979b2"} { - "name": "GraphLowering.compile_to_module", - "ts": 1727975463864454.0, - "args": null, + "name": "GraphLowering.compile_to_fn", + "ts": 1733527490889665.2, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.864000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "2614bec07c124d967f10a9f150623f5a"} +V1206 15:24:50.889000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ba5d98c4ab6385588c824e9ba44d02fe"} { "name": "code_gen", - "ts": 1727975463864454.0, - "args": null, + "ts": 1733527490889960.8, + "args": { + "fn_name": "GraphLowering.compile_to_module", + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.868000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c941c5142cd262302a4ea27ba7d7eda9"} +V1206 15:24:50.890000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "d2916632a86a0b07c4acc6705aa100b6"} { - "name": "Scheduler.__init__", - "ts": 1727975463868177.8, - "args": null, + "name": "GraphLowering.codegen", + "ts": 1733527490890197.0, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.871000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "19fd40416e912c56022b05ddd45c6f91"} +V1206 15:24:50.891000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "e0af71ca8453aaa26c1d6ae93f92a7e0"} { "name": "Scheduler.__init__", - "ts": 1727975463871282.5, + "ts": 1733527490891018.5, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.892000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "6c816233fc4897d4d8763de32b6039c5"} + { + "name": "Scheduler.fused_nodes", + "ts": 1733527490892516.8, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.892000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "398867777111a34eaebfec44bc9bc0e0"} + { + "name": "Scheduler.fused_nodes", + "ts": 1733527490892802.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 } +V1206 15:24:50.895000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "4348580c03d299cf39e06d0f77951243"} + { + "name": "Scheduler.__init__", + "ts": 1733527490895243.5, + "args": { + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.871000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "1ae33988f4240ac670d91f8fe8bc3613"} +V1206 15:24:50.895000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "37369aaefd17f9a63be8fcd31ae1e98c"} { "name": "Scheduler.codegen", - "ts": 1727975463871646.2, - "args": null, + "ts": 1733527490895454.5, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.881000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "4ea7f96a10ffbc186f91b8fbf392a78d"} +V1206 15:24:50.900000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "1b9346463f71bc6d13a662a5be2be40b"} { "name": "Scheduler.codegen", - "ts": 1727975463881456.5, + "ts": 1733527490900442.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.881000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "72cc59f88d66ead0086c546b18c184d7"} +V1206 15:24:50.900000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "54ce92e7f7ba26f1e39c7094e3194d16"} { "name": "PythonWrapperCodegen.generate", - "ts": 1727975463881839.8, - "args": null, + "ts": 1733527490900664.2, + "args": { + "compile_id": "1/0" + }, "ph": "B", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.884000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "20792f3cea34e70b42f7c6d998aaa598"} +V1206 15:24:50.901000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "505184b3d58c91c853c86f14cf8b9246"} { "name": "PythonWrapperCodegen.generate", - "ts": 1727975463884612.0, + "ts": 1733527490901874.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 } +V1206 15:24:50.902000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "fac85554adaad8ec3734fccc5ed505af"} + { + "name": "GraphLowering.codegen", + "ts": 1733527490902086.5, + "args": { + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_code": {"filename": "/tmp/oulgen/tmp4z1i5ywe/si/csitvhfwicmtxv44ng5kkavccd5rcpalvwhjof7rk2hputpzelxm.py"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c453185bb704fad4bd52bdb872abe96c"} +V1206 15:24:50.902000 1667746 torch/_inductor/graph.py:2030] {"inductor_output_code": {"filename": "/tmp/tmpsn77tlsm/gg/cgg6gukzbkegr5nqxvu5c6limjypanrlgt5z3sah5y2lzsjiugp7.py"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "45e11dd29b32835a24d78e4a6ab62a07"} # AOT ID: ['2_inference'] from ctypes import c_void_p, c_long, c_int import torch @@ -4389,7 +5947,14 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c from torch._C import _cuda_getCurrentRawStream as get_raw_stream import triton import triton.language as tl - from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph + from torch._inductor.runtime.triton_heuristics import ( + grid, + split_scan_grid, + grid_combo_kernels, + start_graph, + end_graph, + cooperative_reduction_grid, + ) aten = torch.ops.aten inductor_ops = torch.ops.inductor @@ -4401,14 +5966,15 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor alloc_from_pool = torch.ops.inductor._alloc_from_pool async_compile = AsyncCompile() + empty_strided_p2p = torch._C._distributed_c10d._SymmetricMemory.empty_strided_p2p - # kernel path: /tmp/oulgen/tmp4z1i5ywe/vr/cvrnbiynkuy34l7cucuaotjk5vdulwxhwmvmwz7gy5423p3sv57h.py + # kernel path: /tmp/tmpsn77tlsm/4s/c4sfc3fngwlnbougmdi6kqtjf5vlgrowumdc7a2kkh5gxxav655m.py # Topologically Sorted Source Nodes: [flex_attention], Original ATen: [] # Source node to ATen node mapping: # flex_attention => flex_attention # Graph fragment: - # %flex_attention : [num_users=1] = call_function[target=torch.ops.higher_order.flex_attention](args = (%arg0_1, %arg1_1, %arg2_1, %sdpa_score0, (%arg3_1, %arg4_1, %arg5_1, %arg6_1, %arg7_1, %arg8_1, %arg9_1, %arg10_1, 128, 128, %sdpa_mask0), 0.125, {ROWS_GUARANTEED_SAFE: False, PRESCALE_QK: False, OUTPUT_LOGSUMEXP: False}, (), ()), kwargs = {}) + # %flex_attention : [num_users=1] = call_function[target=torch.ops.higher_order.flex_attention](args = (%arg0_1, %arg1_1, %arg2_1, %sdpa_score0, (%arg3_1, %arg4_1, %arg5_1, %arg6_1, %arg7_1, %arg8_1, %arg9_1, %arg10_1, 128, 128, %sdpa_mask0), 0.125, {PRESCALE_QK: False, ROWS_GUARANTEED_SAFE: False, BLOCKS_ARE_CONTIGUOUS: False, OUTPUT_LOGSUMEXP: True}, (), ()), kwargs = {}) triton_tem_fused_0 = async_compile.triton('triton_tem_fused_0', ''' import triton import triton.language as tl @@ -4416,19 +5982,20 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c from torch._inductor.runtime import triton_helpers, triton_heuristics from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math - from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties + from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, DeviceProperties @triton_heuristics.template( num_stages=3, num_warps=4, - triton_meta={'signature': {'arg_Q': '*fp32', 'arg_K': '*fp32', 'arg_V': '*fp32', 'arg_LSE': '*fp32', 'arg_KV_NUM_BLKS': '*i32', 'arg_KV_IDX': '*i32', 'arg_FULL_KV_NUM_BLKS': '*i32', 'arg_FULL_KV_IDX': '*i32', 'out_ptr0': '*fp32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108, warp_size=32), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]}, - inductor_meta={'kernel_name': 'triton_tem_fused_0', 'backend_hash': 'FB2CA426CF35F271C56C0D69873498391AC248E25890F2B631CA8B52D56952BD', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': False, 'autotune_pointwise': True, 'autotune_remote_cache': False, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}, + triton_meta={'signature': {'arg_Q': '*fp32', 'arg_K': '*fp32', 'arg_V': '*fp32', 'arg_LSE': '*fp32', 'arg_KV_NUM_BLKS': '*i32', 'arg_KV_IDX': '*i32', 'arg_FULL_KV_NUM_BLKS': '*i32', 'arg_FULL_KV_IDX': '*i32', 'out_ptr0': '*fp32'}, 'device': DeviceProperties(type='cuda', index=0, cc=90, major=9, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=132, warp_size=32), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]}, + inductor_meta={'kernel_name': 'triton_tem_fused_0', 'backend_hash': '562E840D41CEB1D8E51DE726EA7592B0C37A0C6FBD72CF4E958863CEC11D41A7', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': False, 'autotune_pointwise': True, 'autotune_remote_cache': False, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}, ) @triton.jit def triton_tem_fused_0(arg_Q, arg_K, arg_V, arg_LSE, arg_KV_NUM_BLKS, arg_KV_IDX, arg_FULL_KV_NUM_BLKS, arg_FULL_KV_IDX, out_ptr0): - ROWS_GUARANTEED_SAFE : tl.constexpr = False PRESCALE_QK : tl.constexpr = False - OUTPUT_LOGSUMEXP : tl.constexpr = False + ROWS_GUARANTEED_SAFE : tl.constexpr = False + BLOCKS_ARE_CONTIGUOUS : tl.constexpr = False + OUTPUT_LOGSUMEXP : tl.constexpr = True FLOAT32_PRECISION : tl.constexpr = 'ieee' IS_DIVISIBLE : tl.constexpr = True SM_SCALE : tl.constexpr = 0.125 @@ -4476,6 +6043,8 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c # about 20% more numerical error, but slightly faster. # ROWS_GUARANTEED_SAFE: Is it guaranteed that at least one value in each row # is not masked out? If so, we can skip an extra safety check + # BLOCKS_ARE_CONTIGUOUS: Is it guaranteed that all blocks in the mask are + # contiguous? If so, we don't need to do an indirect jump for every block tl.static_assert(SPARSE_Q_BLOCK_SIZE >= BLOCK_M and SPARSE_Q_BLOCK_SIZE % BLOCK_M == 0) tl.static_assert(SPARSE_KV_BLOCK_SIZE >= BLOCK_N and SPARSE_KV_BLOCK_SIZE % BLOCK_N == 0) @@ -4641,11 +6210,10 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c idx_d = tl.arange(0, V_HEAD_DIM)[None, :] mask = idx_m < Q_LEN - # TODO generalize and add proper mask support - xindex = idx_d + (64*idx_m) + (32768*idx_hq) + (131072*idx_zq) - tl.store(out_ptr0 + (tl.broadcast_to(idx_d + (64*idx_m) + (32768*idx_hq), acc.shape)), acc, mask) - # TODO dont want to write this if we dont require grad + xindex = idx_d + 64*idx_m + 32768*idx_hq + 131072*idx_zq + tl.store(out_ptr0 + (tl.broadcast_to(idx_d + 64*idx_m + 32768*idx_hq, acc.shape)), acc, mask) + if OUTPUT_LOGSUMEXP: off_hz = tl.program_id(1) l_ptrs = LSE + off_hz * Q_LEN + offs_m @@ -4672,9 +6240,10 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c IS_FULL_BLOCKS, ): # Redefines all kernel parameters (BLOCK_M, etc.) so we don't need to plumb them all through - ROWS_GUARANTEED_SAFE : tl.constexpr = False PRESCALE_QK : tl.constexpr = False - OUTPUT_LOGSUMEXP : tl.constexpr = False + ROWS_GUARANTEED_SAFE : tl.constexpr = False + BLOCKS_ARE_CONTIGUOUS : tl.constexpr = False + OUTPUT_LOGSUMEXP : tl.constexpr = True FLOAT32_PRECISION : tl.constexpr = 'ieee' IS_DIVISIBLE : tl.constexpr = True SM_SCALE : tl.constexpr = 0.125 @@ -4726,7 +6295,7 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c # update pointers offset = get_offset_for_next_block( start_n, kv_indices, kv_num_blocks, - SPARSE_KV_BLOCK_SIZE, SPARSE_KV_MULTIPLE, BLOCK_N + SPARSE_KV_BLOCK_SIZE, SPARSE_KV_MULTIPLE, BLOCK_N, BLOCKS_ARE_CONTIGUOUS ) V_block_ptr = tl.advance(V_block_ptr, (offset, 0)) @@ -4738,13 +6307,18 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c @triton.jit - def get_offset_for_next_block(loop_iter, col_indices, total_blocks, SPARSE_BLOCK, SPARSE_BLOCK_MULTIPLE, BLOCK): + def get_offset_for_next_block( + loop_iter, col_indices, total_blocks, + SPARSE_BLOCK, SPARSE_BLOCK_MULTIPLE, BLOCK, + BLOCKS_ARE_CONTIGUOUS: tl.constexpr + ): + if BLOCKS_ARE_CONTIGUOUS: + return BLOCK cur_block_idx = loop_iter // SPARSE_BLOCK_MULTIPLE cur_block = tl.load(col_indices + cur_block_idx, eviction_policy="evict_last") next_block = tl.load(col_indices + cur_block_idx + 1, eviction_policy="evict_last", mask=cur_block_idx + 1 < total_blocks) needs_jump = (loop_iter + 1) % SPARSE_BLOCK_MULTIPLE == 0 jump_to_block = (next_block - cur_block ) * SPARSE_BLOCK - (SPARSE_BLOCK_MULTIPLE - 1) * BLOCK - offset = jump_to_block * needs_jump + (1 - needs_jump) * BLOCK return offset @@ -4760,9 +6334,10 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c IS_FULL_BLOCKS, CHECK_BLOCK_BOUNDARY=False, ): # Redefines all kernel parameters (BLOCK_M, etc.) so we don't need to plumb them all through - ROWS_GUARANTEED_SAFE : tl.constexpr = False PRESCALE_QK : tl.constexpr = False - OUTPUT_LOGSUMEXP : tl.constexpr = False + ROWS_GUARANTEED_SAFE : tl.constexpr = False + BLOCKS_ARE_CONTIGUOUS : tl.constexpr = False + OUTPUT_LOGSUMEXP : tl.constexpr = True FLOAT32_PRECISION : tl.constexpr = 'ieee' IS_DIVISIBLE : tl.constexpr = True SM_SCALE : tl.constexpr = 0.125 @@ -4809,7 +6384,7 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c if CHECK_BLOCK_BOUNDARY: - mask_mod_output = tl.where(offs_n < KV_LEN, mask_mod_output, float("-inf")) + mask_mod_output = tl.where(offs_n < KV_LEN, mask_mod_output, False) # apply mask for partially unmasked blocks post_mod_scores = tl.where(mask_mod_output, post_mod_scores, float("-inf")) @@ -4847,7 +6422,7 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c return acc, l_i, m_i ''', device_str='cuda') - meta0 = {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False, 'FLOAT32_PRECISION': "'ieee'", 'IS_DIVISIBLE': True, 'SM_SCALE': 0.125, 'GQA_SHARED_HEADS': 1, 'HAS_FULL_BLOCKS': True, 'QK_HEAD_DIM': 64, 'V_HEAD_DIM': 64, 'BLOCK_M': 128, 'BLOCK_N': 32, 'SPARSE_Q_BLOCK_SIZE': 128, 'SPARSE_KV_BLOCK_SIZE': 128} + meta0 = {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True, 'FLOAT32_PRECISION': "'ieee'", 'IS_DIVISIBLE': True, 'SM_SCALE': 0.125, 'GQA_SHARED_HEADS': 1, 'HAS_FULL_BLOCKS': True, 'QK_HEAD_DIM': 64, 'V_HEAD_DIM': 64, 'BLOCK_M': 128, 'BLOCK_N': 32, 'SPARSE_Q_BLOCK_SIZE': 128, 'SPARSE_KV_BLOCK_SIZE': 128} async_compile.wait(globals()) @@ -4907,46 +6482,430 @@ V1003 10:11:03.885000 2235078 torch/_inductor/graph.py:1899] {"inductor_output_c from torch._inductor.wrapper_benchmark import compiled_module_main compiled_module_main('None', benchmark_compiled_module) -V1003 10:11:10.314000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0c6c8a1109a650b2f61d8b8a52500c9f"} +V1206 15:24:50.902000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "7466f2d4d61dfc1690516495f37138a1"} + { + "name": "PyCodeCache.load_by_key_path", + "ts": 1733527490902875.0, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:50.916000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c77ba023033616cfe1f998da2fef4de2"} + { + "name": "async_compile.wait", + "ts": 1733527490916172.8, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:54.126000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "07f6598362ca233fcf839b83bd69ac34"} + { + "name": "async_compile.wait", + "ts": 1733527494126241.8, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:54.126000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "7a197b62b8d8d93579b42512d9f7a21f"} + { + "name": "PyCodeCache.load_by_key_path", + "ts": 1733527494126712.0, + "args": { + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:54.127000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "31dde7ddbbe82ef40947971a84c580bc"} { "name": "code_gen", - "ts": 1727975470313939.2, + "ts": 1733527494127082.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 + "fn_name": "GraphLowering.compile_to_module", + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 } +V1206 15:24:54.127000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "e32bbd039d869c1f2168335fbd92d76f"} + { + "name": "GraphLowering.compile_to_fn", + "ts": 1733527494127503.0, + "args": { + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:10.314000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "3d6e3c947033eed0a9cadd2013455900"} +V1206 15:24:54.128000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "fcdee348cde25167c18c8dcbcf951830"} + { + "name": "TritonBundler.collect", + "ts": 1733527494128406.5, + "args": { + "compile_id": "1/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:24:54.129000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "4c9ec5066284075c872b557d2cf8b5e2"} { - "name": "GraphLowering.compile_to_module", - "ts": 1727975470314540.5, + "name": "TritonBundler.collect", + "ts": 1733527494129814.8, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 + "compile_id": "1/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 } +V1206 15:24:54.134000 1667746 torch/_dynamo/utils.py:1327] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "ab69c518495e835b07c9f8fc02777882"} + { + "name": "fx_graph_cache_miss", + "ts": 1733527490856164.5, + "args": { + "key": "fmjzafxtfmjyhhkl3afwwy3gjat2obuydjrd6epcvjw6l5s2v7yd", + "components": [ + "[osxu7jvxrxvfgwz7co25iilkcoi7r6wwkfxjkdf7i35nrmqqnas] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[3]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[4]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[5]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[6]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[7]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[8]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", + "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", + "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", + "[kcuxe2zwm3mzv2uk6adm6iskoy35bqfv725twacrdewod2dbl5d] inputs_to_check[3]: 3", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inputs_to_check[4]: 4", + "[qs5hilycp4ew4ivtc7m5jaxp7q4pm5slioxw3fi3ur6ei65ybz4] inputs_to_check[5]: 5", + "[agkvbkaha53nbz3aeeuhvxjvvc4glhfjofzkg6g2qjoo2e5otcx] inputs_to_check[6]: 6", + "[j3s5elu6itwgjafc7rzhy4whrbufl6kfmlufjhh25grt643bk5f] inputs_to_check[7]: 7", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inputs_to_check[8]: 8", + "[qlgfiyqewrmkgqth2qm6wkq2ja5lzkapg3ypgnvoyfqqnidaoj3] inputs_to_check[9]: 9", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", + "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", + "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[bundled_autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", + "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", + "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", + "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", + "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", + "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", + "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", + "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" + ], + "cache_event_time": 1733527490862706044, + "cache_state": "miss", + "triton_bundler_meta": "TritonBundlerMetadata(cached_kernel_names=['triton_tem_fused_0'])", + "time_taken_ns": 3272216042, + "compile_id": "1/0" }, - "ph": "E", + "ph": "i", "cat": "dynamo_timed", "tid": 0, - "pid": 0 + "pid": 0, + "s": "p" } -V1003 10:11:10.316000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "17b87920e3cd4803f6a7235280e07423"} +V1206 15:24:54.135000 1667746 torch/_inductor/compile_fx.py:751] {"artifact": {"name": "fx_graph_cache_miss", "encoding": "json"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "c9c893906731ce022aed26f875fcd820"} + {"key": "fmjzafxtfmjyhhkl3afwwy3gjat2obuydjrd6epcvjw6l5s2v7yd", "components": ["[osxu7jvxrxvfgwz7co25iilkcoi7r6wwkfxjkdf7i35nrmqqnas] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[3]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[4]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[5]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[6]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[7]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[8]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", "[kcuxe2zwm3mzv2uk6adm6iskoy35bqfv725twacrdewod2dbl5d] inputs_to_check[3]: 3", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inputs_to_check[4]: 4", "[qs5hilycp4ew4ivtc7m5jaxp7q4pm5slioxw3fi3ur6ei65ybz4] inputs_to_check[5]: 5", "[agkvbkaha53nbz3aeeuhvxjvvc4glhfjofzkg6g2qjoo2e5otcx] inputs_to_check[6]: 6", "[j3s5elu6itwgjafc7rzhy4whrbufl6kfmlufjhh25grt643bk5f] inputs_to_check[7]: 7", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inputs_to_check[8]: 8", "[qlgfiyqewrmkgqth2qm6wkq2ja5lzkapg3ypgnvoyfqqnidaoj3] inputs_to_check[9]: 9", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[bundled_autotune_remote_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None"], "cache_event_time": 1733527490862706044, "cache_state": "miss", "triton_bundler_meta": "TritonBundlerMetadata(cached_kernel_names=['triton_tem_fused_0'])", "time_taken_ns": 3272216042, "compile_id": "1/0"} +V1206 15:24:54.135000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "8448954b82f0bb83dace88546d83f2e6"} { - "name": "fx_graph_cache_miss", - "ts": 1727975463815134.0, + "name": "inductor_compile", + "ts": 1733527494135534.2, "args": { - "key": "flfpwf422lzhb6yszystfjdqdgiiqyds6rdbqc57o3a6n4uehjgp", + "fn_name": "compile_fx_inner", + "compile_id": "1/0", + "is_backward": false, + "cache_state": "miss", + "cache_event_time": 1733527490856164558, + "key": "fmjzafxtfmjyhhkl3afwwy3gjat2obuydjrd6epcvjw6l5s2v7yd", "components": [ - "[nhtxa6qinb75ty5rqjx3bdkokm2s3bz3nrhke3zmvm73qypm2md] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[osxu7jvxrxvfgwz7co25iilkcoi7r6wwkfxjkdf7i35nrmqqnas] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'PRESCALE_QK': False, 'ROWS_GUARANTEED_SAFE': False, 'BLOCKS_ARE_CONTIGUOUS': False, 'OUTPUT_LOGSUMEXP': True}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", @@ -4959,6 +6918,7 @@ V1003 10:11:10.316000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", + "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", @@ -4966,7 +6926,6 @@ V1003 10:11:10.316000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", - "[f44ag5aflby2bkxl7a4k6whljrk7jat7bmreuxklei4p3czhk7p] fx_kwargs[user_visible_outputs]: {'getitem': None}", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", @@ -4980,396 +6939,379 @@ V1003 10:11:10.316000 2235078 torch/_dynamo/utils.py:1020] {"chromium_event": {} "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", - "[7uhqwjfn75ek3woo3k7em2mluon5hx2ojvzlevlvjvz6xfxjhzl] torch_version: ", - "[c3z7bmoxyo6gl5hi47v6dc7jwsl55b3asd75nr25uyengi5ah3p] system_info[device]: {'name': 'NVIDIA PG509-210'}", - "[3fb7kae6ogkdd4zcm3fkjoipdpybxhn4aoxzv7z7xsfwq233e4l] system_info[version]: {'triton': '3.1.0+5fe38ffd73dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-a8fb7be728d460b7ec64ab62edb8af1bbca8994fd718cde7178d46bad64530a1-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.0'}", - "[z5x2bdhir5lzlbti73vdbfulnuu5vinzpwgmmgf4rjb775tzl3h] system_info[hash]: 9698c97edde4a99a2f3b54bbd0db5291bbcdb75c83acb376ccff61fb0bf0ac1a", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[abi_compatible]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_dump_consts_bin]: False", - "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", - "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", - "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", - "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", - "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", - "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", - "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", - "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", - "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", - "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", - "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", - "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", - "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", - "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cuda.generate_test_runner]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", - "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[bundled_autotune_remote_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[enable_auto_functionalized_v2]: False", - "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", - "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", - "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", - "[62lrdx35b7hnumwb7mp5oc5y5csm2abylvtdzfloct3noaqov3n] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'post_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", - "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", - "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", - "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", - "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", - "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", - "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", + "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", - "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", - "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", - "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", - "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", - "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", - "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", - "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", - "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_post_pass]: None", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_pre_pass]: None", - "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", - "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", - "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", - "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", - "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", - "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_peak_memory]: False", - "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", - "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", - "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", - "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", + "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", - "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", - "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", - "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", - "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", - "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", - "[vzzema5ityqj2wepdmkulue7q5pcevdr5h27oxxutf35d4tjume] inductor_config[triton_kernel_default_layout_constraint]: flexible_layout", - "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", - "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", - "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", - "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False" + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", + "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" ], - "cache_event_time": 1727975463815134085, - "cache_state": "miss", - "time_taken_ns": 6500727742 - }, - "ph": "i", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0, - "s": "p" - } -V1003 10:11:10.317000 2235078 torch/_inductor/codecache.py:1463] {"artifact": {"name": "fx_graph_cache_miss", "encoding": "json"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "2b8d4494cf5a9a8e000f402106cf2ba5"} - {"key": "flfpwf422lzhb6yszystfjdqdgiiqyds6rdbqc57o3a6n4uehjgp", "components": ["[nhtxa6qinb75ty5rqjx3bdkokm2s3bz3nrhke3zmvm73qypm2md] gm: (\n (sdpa_score0): ()\n (sdpa_mask0): ()\n)\n\n\n\ndef forward(self, arg0_1, arg1_1, arg2_1, arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1):\n sdpa_score0 = self.sdpa_score0\n sdpa_mask0 = self.sdpa_mask0\n flex_attention = torch.ops.higher_order.flex_attention(arg0_1, arg1_1, arg2_1, sdpa_score0, (arg3_1, arg4_1, arg5_1, arg6_1, arg7_1, arg8_1, arg9_1, arg10_1, 128, 128, sdpa_mask0), 0.125, {'ROWS_GUARANTEED_SAFE': False, 'PRESCALE_QK': False, 'OUTPUT_LOGSUMEXP': False}, (), ()); arg0_1 = arg1_1 = arg2_1 = sdpa_score0 = arg3_1 = arg4_1 = arg5_1 = arg6_1 = arg7_1 = arg8_1 = arg9_1 = arg10_1 = sdpa_mask0 = None\n getitem = flex_attention[0]; flex_attention = None\n return (getitem,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[1]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[avf2u3luxvyabchjhbddapcjn5gev47wfdtkrprayuhv6lf2z6u] example_inputs[2]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 4, 512, 64]), stride=(131072, 32768, 64, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[3]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[4]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[5]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[6]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[7]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[8]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[zsk3gejenkcvvwhiyk36u5zdnlrcs6wgy3pina3csuierfd2zri] example_inputs[9]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16]), stride=(16, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[hnbjjzmb63q27mbr22eubaelyb423burv27meouma6ccysmwu6g] example_inputs[10]: TensorMetadata(dtype=torch.int32, shape=torch.Size([1, 1, 16, 16]), stride=(256, 256, 16, 1), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.contiguous_format, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", "[f44ag5aflby2bkxl7a4k6whljrk7jat7bmreuxklei4p3czhk7p] fx_kwargs[user_visible_outputs]: {'getitem': None}", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inputs_to_check[1]: 1", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inputs_to_check[2]: 2", "[kcuxe2zwm3mzv2uk6adm6iskoy35bqfv725twacrdewod2dbl5d] inputs_to_check[3]: 3", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inputs_to_check[4]: 4", "[qs5hilycp4ew4ivtc7m5jaxp7q4pm5slioxw3fi3ur6ei65ybz4] inputs_to_check[5]: 5", "[agkvbkaha53nbz3aeeuhvxjvvc4glhfjofzkg6g2qjoo2e5otcx] inputs_to_check[6]: 6", "[j3s5elu6itwgjafc7rzhy4whrbufl6kfmlufjhh25grt643bk5f] inputs_to_check[7]: 7", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inputs_to_check[8]: 8", "[qlgfiyqewrmkgqth2qm6wkq2ja5lzkapg3ypgnvoyfqqnidaoj3] inputs_to_check[9]: 9", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inputs_to_check[10]: 10", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", "[7uhqwjfn75ek3woo3k7em2mluon5hx2ojvzlevlvjvz6xfxjhzl] torch_version: ", "[c3z7bmoxyo6gl5hi47v6dc7jwsl55b3asd75nr25uyengi5ah3p] system_info[device]: {'name': 'NVIDIA PG509-210'}", "[3fb7kae6ogkdd4zcm3fkjoipdpybxhn4aoxzv7z7xsfwq233e4l] system_info[version]: {'triton': '3.1.0+5fe38ffd73dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-a8fb7be728d460b7ec64ab62edb8af1bbca8994fd718cde7178d46bad64530a1-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.0'}", "[z5x2bdhir5lzlbti73vdbfulnuu5vinzpwgmmgf4rjb775tzl3h] system_info[hash]: 9698c97edde4a99a2f3b54bbd0db5291bbcdb75c83acb376ccff61fb0bf0ac1a", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[abi_compatible]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_dump_consts_bin]: False", "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_local_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_remote_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", "[g7rrnbg5yonzux3cfj5ovre5lob3ayda7qcfpxjvtwmiz4uicii] inductor_config[cpp.min_chunk_size]: 4096", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cuda.generate_test_runner]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_index_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[enable_auto_functionalized_v2]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fx_graph_remote_cache]: False", "[62lrdx35b7hnumwb7mp5oc5y5csm2abylvtdzfloct3noaqov3n] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'post_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[generate_intermediate_hooks]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[post_grad_custom_pre_pass]: None", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_peak_memory]: False", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_pointwise]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", "[vzzema5ityqj2wepdmkulue7q5pcevdr5h27oxxutf35d4tjume] inductor_config[triton_kernel_default_layout_constraint]: flexible_layout", "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False"], "cache_event_time": 1727975463815134085, "cache_state": "miss", "time_taken_ns": 6500727742} -V1003 10:11:10.318000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "59401c365876e3190b828c9baef54d73"} - { - "name": "inductor_compile", - "ts": 1727975470318146.5, - "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } - }, - "ph": "E", - "cat": "dynamo_timed", - "tid": 0, - "pid": 0 - } -V1003 10:11:10.318000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "75e1054f04f269f6411028767770c53c"} - { - "name": "compile_fx_inner", - "ts": 1727975470318439.5, - "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } + "cache_bypass_reason": null, + "remote_cache_enabled": false, + "local_cache_enabled": true }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:10.319000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "54543856a0e4177bd66c38146fda7120"} +V1206 15:24:54.136000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "a9d90840e9080e38cc5cc8f43d64b9e7"} { "name": "compile_fx..fw_compiler_base", - "ts": 1727975470319104.2, + "ts": 1733527494136361.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:10.322000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0017891a0ab9c15e62e1b8125198b31d"} +V1206 15:24:54.138000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "9174f1070d6248c4511951edfe3f1dd1"} { "name": "create_aot_dispatcher_function", - "ts": 1727975470322488.8, + "ts": 1733527494138570.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:10.323000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0927305357c25444fda9907c2d33a10f"} +V1206 15:24:54.138000 1667746 torch/_dynamo/utils.py:1327] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "32e73583977fb2c6cb9b34dd32b9a8f7"} { - "name": "backend_compile", - "ts": 1727975470323094.0, + "name": "autograd_cache_bypass", + "ts": 1733527490809971.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } + "cache_bypass_reason": "Unsupported call_function target flex_attention. \n Function module: torch.ops.higher_order, \nFunction name: flex_attention", + "cache_bypass_hard_exception": false, + "key": null, + "cache_state": "bypass", + "components": [], + "compile_id": "1/0" }, - "ph": "E", + "ph": "i", "cat": "dynamo_timed", "tid": 0, - "pid": 0 + "pid": 0, + "s": "p" } -V1003 10:11:10.323000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f3906ae17fd7aa60928cc5980667652c"} +V1206 15:24:54.138000 1667746 torch/_functorch/_aot_autograd/autograd_cache.py:763] {"artifact": {"name": "aotautograd_cache_hash", "encoding": "json"}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "5601d02186053adcc1ba29fd248c1d20"} + {"cache_bypass_reason": "Unsupported call_function target flex_attention. \n Function module: torch.ops.higher_order, \nFunction name: flex_attention", "cache_bypass_hard_exception": false, "key": null, "cache_state": "bypass", "components": [], "compile_id": "1/0"} +V1206 15:24:54.139000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "cca36f8b4f9f5ea8ad866b857a634eb8"} { - "name": "OutputGraph.call_user_compiler", - "ts": 1727975470323375.2, + "name": "backend_compile", + "ts": 1733527494139170.2, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "1/0", + "requires_subclass_dispatch": false, + "dispatch_mode": "inference", + "cache_state": "bypass", + "cache_event_time": 1733527490809971264, + "key": null, + "components": [], + "cache_bypass_reason": "Unsupported call_function target flex_attention. \n Function module: torch.ops.higher_order, \nFunction name: flex_attention", + "remote_cache_enabled": false, + "local_cache_enabled": true }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:10.354000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_str": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "1b9c89081ea4384ef6cc3f1c7563841f"} +V1206 15:24:54.153000 1667746 torch/_dynamo/guards.py:2315] {"dynamo_cpp_guards_str": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "b9091c16617d57e52682988042d2b9ea"} TREE_GUARD_MANAGER: +- RootGuardManager - | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:471 in init_ambient_guards + | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:484 in init_ambient_guards | +- GLOBAL_STATE: ___check_global_state() | +- TORCH_FUNCTION_MODE_STACK: ___check_torch_function_mode_stack() - | +- GuardManager: source=L['k'], accessed_by=DictGetItemGuardAccessor(k) - | | +- TYPE_MATCH: ___check_type_id(L['k'], 82028112) + | +- GuardManager: source=L['k'], accessed_by=DictGetItemGuardAccessor('k') + | | +- TYPE_MATCH: ___check_type_id(L['k'], 82181376) | | +- TENSOR_MATCH: check_tensor(L['k'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) | | +- NO_HASATTR: hasattr(L['k'], '_dynamo_dynamic_indices') == False | | +- NO_TENSOR_ALIASING: check_no_aliasing(L['k'], L['q'], L['v'], L['block_mask'].q_indices, L['block_mask'].kv_indices, L['block_mask'].q_num_blocks, L['block_mask'].kv_num_blocks, L['block_mask'].full_q_indices, L['block_mask'].full_kv_indices, L['block_mask'].full_q_num_blocks, L['block_mask'].full_kv_num_blocks) - | +- GuardManager: source=L['q'], accessed_by=DictGetItemGuardAccessor(q) - | | +- TYPE_MATCH: ___check_type_id(L['q'], 82028112) + | | +- GuardManager: source=L['k'].dim, accessed_by=GetAttrGuardAccessor(dim) + | | +- GuardManager: source=L['k'].size, accessed_by=GetAttrGuardAccessor(size) + | +- GuardManager: source=L['q'], accessed_by=DictGetItemGuardAccessor('q') + | | +- TYPE_MATCH: ___check_type_id(L['q'], 82181376) | | +- TENSOR_MATCH: check_tensor(L['q'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) | | +- NO_HASATTR: hasattr(L['q'], '_dynamo_dynamic_indices') == False | | +- NO_TENSOR_ALIASING - | +- GuardManager: source=L['v'], accessed_by=DictGetItemGuardAccessor(v) - | | +- TYPE_MATCH: ___check_type_id(L['v'], 82028112) + | | +- GuardManager: source=L['q'].dim, accessed_by=GetAttrGuardAccessor(dim) + | | +- GuardManager: source=L['q'].size, accessed_by=GetAttrGuardAccessor(size) + | +- GuardManager: source=L['v'], accessed_by=DictGetItemGuardAccessor('v') + | | +- TYPE_MATCH: ___check_type_id(L['v'], 82181376) | | +- TENSOR_MATCH: check_tensor(L['v'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 4, 512, 64], stride=[131072, 32768, 64, 1]) | | +- NO_HASATTR: hasattr(L['v'], '_dynamo_dynamic_indices') == False | | +- NO_TENSOR_ALIASING - | +- GuardManager: source=L['block_mask'], accessed_by=DictGetItemGuardAccessor(block_mask) - | | +- TYPE_MATCH: ___check_type_id(L['block_mask'], 387600320) + | | +- GuardManager: source=L['v'].dim, accessed_by=GetAttrGuardAccessor(dim) + | | +- GuardManager: source=L['v'].size, accessed_by=GetAttrGuardAccessor(size) + | +- GuardManager: source=L['block_mask'], accessed_by=DictGetItemGuardAccessor('block_mask') + | | +- TYPE_MATCH: ___check_type_id(L['block_mask'], 139679664) | | +- GuardManager: source=L['block_mask'].mask_mod, accessed_by=GetAttrGuardAccessor(mask_mod) | | | +- GuardManager: source=L['block_mask'].mask_mod.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].mask_mod.__code__, 140413271880128) + | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].mask_mod.__code__, 140062268556768) | | +- GuardManager: source=L['block_mask'].q_indices, accessed_by=GetAttrGuardAccessor(q_indices) | | | +- TENSOR_MATCH: check_tensor(L['block_mask'].q_indices, Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.int32, device=0, requires_grad=False, size=[1, 1, 16, 16], stride=[256, 256, 16, 1]) | | | +- NO_HASATTR: hasattr(L['block_mask'].q_indices, '_dynamo_dynamic_indices') == False | | | +- NO_TENSOR_ALIASING | | +- GuardManager: source=L['block_mask'].BLOCK_SIZE, accessed_by=GetAttrGuardAccessor(BLOCK_SIZE) - | | | +- TYPE_MATCH: ___check_type_id(L['block_mask'].BLOCK_SIZE, 8815232) + | | | +- TYPE_MATCH: ___check_type_id(L['block_mask'].BLOCK_SIZE, 8812224) | | | +- LENGTH_CHECK: len(L['block_mask'].BLOCK_SIZE) == 2 | | | +- GuardManager: source=L['block_mask'].BLOCK_SIZE[0], accessed_by=TupleGetItemGuardAccessor(0) | | | | +- EQUALS_MATCH: L['block_mask'].BLOCK_SIZE[0] == 128 @@ -5406,51 +7348,57 @@ V1003 10:11:10.354000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | +- GuardManager: source=L['block_mask'].as_tuple, accessed_by=GetAttrGuardAccessor(as_tuple) | | | +- GuardManager: source=L['block_mask'].as_tuple, accessed_by=FuncDefaultsGuardAccessor | | | | +- GuardManager: source=L['block_mask'].as_tuple.__defaults__[0], accessed_by=GetItemGuardAccessor(0) - | | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].as_tuple.__defaults__[0], 8911040) - | +- GuardManager: source=L['score_mod2'], accessed_by=DictGetItemGuardAccessor(score_mod2) + | | | | | +- ID_MATCH: ___check_obj_id(L['block_mask'].as_tuple.__defaults__[0], 8908032) + | +- GuardManager: source=L['score_mod2'], accessed_by=DictGetItemGuardAccessor('score_mod2') | | +- GuardManager: source=L['score_mod2'].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | +- ID_MATCH: ___check_obj_id(L['score_mod2'].__code__, 140413275216112) - | +- GuardManager: source=L['flex_attention'], accessed_by=DictGetItemGuardAccessor(flex_attention) + | | | +- ID_MATCH: ___check_obj_id(L['score_mod2'].__code__, 140062269255152) + | +- GuardManager: source=L['flex_attention'], accessed_by=DictGetItemGuardAccessor('flex_attention') | | +- GuardManager: source=L['flex_attention'].__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__code__, 387082992) + | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__code__, 139318784) | | +- GuardManager: source=L['flex_attention'], accessed_by=FuncDefaultsGuardAccessor | | | +- GuardManager: source=L['flex_attention'].__defaults__[2], accessed_by=GetItemGuardAccessor(2) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[2], 8825760) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[2], 8822752) | | | +- GuardManager: source=L['flex_attention'].__defaults__[3], accessed_by=GetItemGuardAccessor(3) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[3], 8910592) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[3], 8907584) | | | +- GuardManager: source=L['flex_attention'].__defaults__[4], accessed_by=GetItemGuardAccessor(4) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[4], 8910592) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[4], 8907584) | | | +- GuardManager: source=L['flex_attention'].__defaults__[5], accessed_by=GetItemGuardAccessor(5) - | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[5], 8825760) + | | | | +- ID_MATCH: ___check_obj_id(L['flex_attention'].__defaults__[5], 8822752) | +- GuardManager: source=G, accessed_by=GlobalsGuardAccessor - | | +- GuardManager: source=G['__builtins_dict___10'], accessed_by=DictGetItemGuardAccessor(__builtins_dict___10) - | | | +- GuardManager: source=G['__builtins_dict___10']['len'], accessed_by=DictGetItemGuardAccessor(len) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['len'], 140413275558816) - | | | +- GuardManager: source=G['__builtins_dict___10']['sum'], accessed_by=DictGetItemGuardAccessor(sum) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['sum'], 140413275559936) - | | | +- GuardManager: source=G['__builtins_dict___10']['list'], accessed_by=DictGetItemGuardAccessor(list) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['list'], 8844320) - | | | +- GuardManager: source=G['__builtins_dict___10']['type'], accessed_by=DictGetItemGuardAccessor(type) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['type'], 8813248) - | | | +- GuardManager: source=G['__builtins_dict___10']['tuple'], accessed_by=DictGetItemGuardAccessor(tuple) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['tuple'], 8815232) - | | | +- GuardManager: source=G['__builtins_dict___10']['object'], accessed_by=DictGetItemGuardAccessor(object) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['object'], 8813984) - | | | +- GuardManager: source=G['__builtins_dict___10']['isinstance'], accessed_by=DictGetItemGuardAccessor(isinstance) - | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['isinstance'], 140413275558496) - | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_utils_dot__pytree) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140411217627952) + | | +- GuardManager: source=G['_140062263790704_c5'], accessed_by=DictGetItemGuardAccessor('_140062263790704_c5') + | | | +- GuardManager: source=G['_140062263790704_c5'].Tensor, accessed_by=GetAttrGuardAccessor(Tensor) + | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c5'].Tensor, 82181376) + | | | | +- GuardManager: source=G['_140062263790704_c5'].Tensor.__bases__, accessed_by=GetAttrGuardAccessor(__bases__) + | | | | | +- GuardManager: source=G['_140062263790704_c5'].Tensor.__bases__[0], accessed_by=TupleGetItemGuardAccessor(0) + | | | | | | +- ID_MATCH: ___check_obj_id(G['_140062263790704_c5'].Tensor.__bases__[0], 140062119703136) + | | +- GuardManager: source=G['__builtins_dict___10'], accessed_by=DictGetItemGuardAccessor('__builtins_dict___10') + | | | +- GuardManager: source=G['__builtins_dict___10']['len'], accessed_by=DictGetItemGuardAccessor('len') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['len'], 140062269592480) + | | | +- GuardManager: source=G['__builtins_dict___10']['sum'], accessed_by=DictGetItemGuardAccessor('sum') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['sum'], 140062269593600) + | | | +- GuardManager: source=G['__builtins_dict___10']['list'], accessed_by=DictGetItemGuardAccessor('list') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['list'], 8841312) + | | | +- GuardManager: source=G['__builtins_dict___10']['type'], accessed_by=DictGetItemGuardAccessor('type') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['type'], 8810240) + | | | +- GuardManager: source=G['__builtins_dict___10']['tuple'], accessed_by=DictGetItemGuardAccessor('tuple') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['tuple'], 8812224) + | | | +- GuardManager: source=G['__builtins_dict___10']['object'], accessed_by=DictGetItemGuardAccessor('object') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['object'], 8810976) + | | | +- GuardManager: source=G['__builtins_dict___10']['isinstance'], accessed_by=DictGetItemGuardAccessor('isinstance') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___10']['isinstance'], 140062269592160) + | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_utils_dot__pytree') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'], 140057600329744) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].TreeSpec, accessed_by=GetAttrGuardAccessor(TreeSpec) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 84866496) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].TreeSpec, 86791872) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf, accessed_by=GetAttrGuardAccessor(_is_leaf) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140411217262720) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_leaf.__code__, 140057601749808) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, accessed_by=GetAttrGuardAccessor(_LEAF_SPEC) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 85171104) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC, 87025712) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, accessed_by=GetAttrGuardAccessor(type) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.type, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, accessed_by=GetAttrGuardAccessor(context) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8825760) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.context, 8822752) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes, accessed_by=GetAttrGuardAccessor(num_nodes) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_nodes == 1 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_leaves, accessed_by=GetAttrGuardAccessor(num_leaves) @@ -5458,37 +7406,37 @@ V1003 10:11:10.354000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children, accessed_by=GetAttrGuardAccessor(num_children) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.num_children == 0 | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, accessed_by=GetAttrGuardAccessor(children_specs) - | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8844320) + | | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs, 8841312) | | | | | +- LENGTH_CHECK: not G['__import_torch_dot_utils_dot__pytree']._LEAF_SPEC.children_specs | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type, accessed_by=GetAttrGuardAccessor(_get_node_type) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140411217262448) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._get_node_type.__code__, 140057601749536) | | | +- DictGuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES, accessed_by=GetAttrGuardAccessor(SUPPORTED_NODES) - | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 519596 + | | | | +- DICT_VERSION: ___dict_version(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES) == 366913 | | | | +- KeyValueManager pair at index=1 | | | | | +- ValueManager: GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]] | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn, accessed_by=GetAttrGuardAccessor(flatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140411196281984) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].flatten_fn.__code__, 140057599989696) | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn, accessed_by=GetAttrGuardAccessor(unflatten_fn) | | | | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140411217182288) + | | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES[list(G['__import_torch_dot_utils_dot__pytree'].SUPPORTED_NODES.keys())[1]].unflatten_fn.__code__, 140057600267792) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper, accessed_by=GetAttrGuardAccessor(_tree_flatten_helper) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140411217413040) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._tree_flatten_helper.__code__, 140057600001392) | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance, accessed_by=GetAttrGuardAccessor(_is_namedtuple_instance) | | | | +- GuardManager: source=G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140411217412592) - | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_comptime'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot__dynamo_dot_comptime) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_comptime'], 140410226912176) - | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_decorators'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot__dynamo_dot_decorators) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'], 140410226910096) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_utils_dot__pytree']._is_namedtuple_instance.__code__, 140057600000944) + | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_comptime'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot__dynamo_dot_comptime') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_comptime'], 140057359527872) + | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_decorators'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot__dynamo_dot_decorators') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'], 140057359526192) | | | +- GuardManager: source=G['__import_torch_dot__dynamo_dot_decorators'].is_compiling, accessed_by=GetAttrGuardAccessor(is_compiling) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'].is_compiling, 140410376252096) - | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_nn_dot_attention_dot__utils) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot__utils'], 140409673896784) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot__dynamo_dot_decorators'].is_compiling, 140057422814624) + | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_nn_dot_attention_dot__utils') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot__utils'], 140048551571072) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS, accessed_by=GetAttrGuardAccessor(_SUPPORTED_HEAD_DIMS) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS, 8844320) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS, 8841312) | | | | +- LENGTH_CHECK: len(G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS) == 10 | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[0], accessed_by=ListGetItemGuardAccessor(0) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[0] == 2 @@ -5502,73 +7450,92 @@ V1003 10:11:10.354000 2235078 torch/_dynamo/guards.py:2311] {"dynamo_cpp_guards_ | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[4] == 32 | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[5], accessed_by=ListGetItemGuardAccessor(5) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot__utils']._SUPPORTED_HEAD_DIMS[5] == 64 - | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], accessed_by=DictGetItemGuardAccessor(__import_torch_dot_nn_dot_attention_dot_flex_attention) - | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], 140409673895824) + | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_nn_dot_attention_dot_flex_attention') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'], 140048551568912) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math, accessed_by=GetAttrGuardAccessor(math) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math, 140413266939392) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math, 140062267404384) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math.sqrt, accessed_by=GetAttrGuardAccessor(sqrt) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math.sqrt, 140413266943072) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].math.sqrt, 140062267408144) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch, accessed_by=GetAttrGuardAccessor(torch) - | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch, 140413267918368) + | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch, 140062263790704) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo, accessed_by=GetAttrGuardAccessor(_dynamo) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo, 140413260098400) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo, 140062262057760) | | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static, accessed_by=GetAttrGuardAccessor(mark_static) | | | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static.__code__, 123166432) + | | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch._dynamo.mark_static.__code__, 125680000) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler, accessed_by=GetAttrGuardAccessor(compiler) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler, 140410826010400) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler, 140057571877776) | | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler.is_dynamo_compiling, accessed_by=GetAttrGuardAccessor(is_dynamo_compiling) - | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler.is_dynamo_compiling, 140410826132992) + | | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.compiler.is_dynamo_compiling, 140057572232544) + | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.is_grad_enabled, accessed_by=GetAttrGuardAccessor(is_grad_enabled) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].torch.is_grad_enabled, 140062253115936) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device, accessed_by=GetAttrGuardAccessor(_validate_device) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device.__code__, 140409673611088) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_device.__code__, 140048552018544) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop, accessed_by=GetAttrGuardAccessor(flex_attention_hop) - | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop, 96992544) + | | | | +- TYPE_MATCH: ___check_type_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop, 96230624) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop.__name__, accessed_by=GetAttrGuardAccessor(__name__) | | | | | +- EQUALS_MATCH: G['__import_torch_dot_nn_dot_attention_dot_flex_attention'].flex_attention_hop.__name__ == 'flex_attention' | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim, accessed_by=GetAttrGuardAccessor(_supported_head_dim) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim.__code__, 140409673231376) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._supported_head_dim.__code__, 140048552864992) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim, accessed_by=GetAttrGuardAccessor(_validate_embed_dim) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim.__code__, 388086512) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_embed_dim.__code__, 139781872) + | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_nestedness, accessed_by=GetAttrGuardAccessor(_validate_nestedness) + | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_nestedness.__code__, accessed_by=GetAttrGuardAccessor(__code__) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_nestedness.__code__, 140048553100560) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input, accessed_by=GetAttrGuardAccessor(_validate_sdpa_input) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input.__code__, 387915104) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._validate_sdpa_input.__code__, 139823744) | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options, accessed_by=GetAttrGuardAccessor(_apply_kernel_options) | | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options.__code__, accessed_by=GetAttrGuardAccessor(__code__) - | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options.__code__, 140409683680752) + | | | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_attention_dot_flex_attention']._apply_kernel_options.__code__, 140048554555888) -V1003 10:11:10.355000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "68cb1c0ec8488a404219d0a05ec80df9"} +V1206 15:24:54.154000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "f5553f623cc33ff3da54884bffba4841"} { "name": "entire_frame_compile", - "ts": 1727975470355549.8, + "ts": 1733527494154160.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } + "fn_name": "_compile.compile_inner", + "compile_id": "1/0" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 } -V1003 10:11:10.355000 2235078 torch/_dynamo/utils.py:988] {"chromium_event": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "eb02aec5687f6fb5c0049bd4411b1b24"} +V1206 15:24:54.156000 1667746 torch/_dynamo/utils.py:1045] {"compilation_metrics": {"pre_grad_pass_time_us": 526, "joint_graph_pass_time_us": 1028, "feature_usage": {"pytorch/remote_cache:fx_graph_memcache_version": true, "pytorch/remote_cache:bundle_triton_into_fx_graph_cache_v2": true}, "post_grad_pass_time_us": 1058, "inductor_code_gen_cumulative_compile_time_us": 3237121, "inductor_cumulative_compile_time_us": 3280008, "aot_autograd_cumulative_compile_time_us": 3332164, "dynamo_cumulative_compile_time_us": 3438591, "frame_key": "2", "co_name": "fn2", "co_filename": "/data/users/xmfan/a/pytorch/test/inductor/test_codecache.py", "co_firstlineno": 465, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 91, "shape_env_guard_count": 0, "graph_op_count": 2, "graph_node_count": 16, "graph_input_count": 11, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": false, "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}", "is_forward": true, "dynamo_compile_time_before_restart_us": 0}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} +V1206 15:24:54.156000 1667746 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 1, "frame_compile_id": 0, "attempt": 0, "has_payload": "0fdda1350847e58c2900e8c2a60a9a2b"} { - "name": "_compile.compile_inner", - "ts": 1727975470355876.5, + "name": "dynamo", + "ts": 1733527494156938.5, "args": { - "cache_stats": { - "fxgraph_cache_hit": 1, - "fxgraph_cache_miss": 2, - "fxgraph_cache_bypass": 0 - } + "compile_id": "1/0", + "frame_key": "2", + "co_name": "fn2", + "co_filename": "/data/users/xmfan/a/pytorch/test/inductor/test_codecache.py", + "co_firstlineno": 465, + "cache_size": 0, + "accumulated_cache_size": 0, + "guard_count": 91, + "shape_env_guard_count": 0, + "graph_op_count": 2, + "graph_node_count": 16, + "graph_input_count": 11, + "fail_type": null, + "fail_reason": null, + "fail_user_frame_filename": null, + "fail_user_frame_lineno": null, + "non_compliant_ops": [], + "compliant_custom_ops": [], + "restart_reasons": [], + "dynamo_time_before_restart_s": 0.0, + "has_guarded_code": true, + "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._decomp\", \"torch.testing\", \"torch._prims\", \"torch._refs\", \"torch.distributions\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": true, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}" }, "ph": "E", "cat": "dynamo_timed", "tid": 0, "pid": 0 - } -V1003 10:11:10.356000 2235078 torch/_dynamo/utils.py:840] {"compilation_metrics": {"compile_id": "1/0", "frame_key": "2", "co_name": "fn2", "co_filename": "/data/users/oulgen/pytorch/test/inductor/test_codecache.py", "co_firstlineno": 385, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 80, "shape_env_guard_count": 0, "graph_op_count": 11, "graph_node_count": 25, "graph_input_count": 11, "start_time": 1727975463.4963815, "entire_frame_compile_time_s": 6.859049081802368, "backend_compile_time_s": 6.648646593093872, "inductor_compile_time_s": 6.517011404037476, "code_gen_time_s": 6.449295282363892, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "possibly_missed_reinplacing_opportunities": 0, "remote_cache_time_saved_s": 0, "structured_logging_overhead_s": 0.085769178, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": true}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} + } \ No newline at end of file diff --git a/tests/inputs/chromium_events.log b/tests/inputs/chromium_events.log index a2217cd..27d67d8 100644 --- a/tests/inputs/chromium_events.log +++ b/tests/inputs/chromium_events.log @@ -1,339 +1,178 @@ -V0807 09:51:32.403000 2824488 torch/_logging/structured.py:22] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/convert_frame.py", 0]} -V0807 09:51:32.403000 2824488 torch/_logging/structured.py:22] {"str": ["/data/users/jjwu/a/pytorch/test/dynamo/test_misc.py", 1]} -V0807 09:51:32.404000 2824488 torch/_logging/structured.py:22] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/test_case.py", 2]} -V0807 09:51:32.404000 2824488 torch/_logging/structured.py:22] {"str": ["/data/users/jjwu/a/pytorch/torch/testing/_internal/common_utils.py", 3]} -V0807 09:51:32.404000 2824488 torch/_logging/structured.py:22] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/main.py", 4]} -V0807 09:51:32.404000 2824488 torch/_logging/structured.py:22] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/runner.py", 5]} -V0807 09:51:32.404000 2824488 torch/_logging/structured.py:22] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/suite.py", 6]} -V0807 09:51:32.404000 2824488 torch/_logging/structured.py:22] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/case.py", 7]} -V0807 09:51:32.404000 2824488 torch/_logging/structured.py:22] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/eval_frame.py", 8]} -V0807 09:51:32.404000 2824488 torch/_dynamo/convert_frame.py:868] {"dynamo_start": {"stack": [{"line": 11137, "name": "", "filename": 1}, {"line": 38, "name": "run_tests", "filename": 2}, {"line": 1270, "name": "run_tests", "filename": 3}, {"line": 101, "name": "__init__", "filename": 4}, {"line": 271, "name": "runTests", "filename": 4}, {"line": 184, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 650, "name": "__call__", "filename": 7}, {"line": 3051, "name": "run", "filename": 3}, {"line": 3023, "name": "_run_custom", "filename": 3}, {"line": 591, "name": "run", "filename": 7}, {"line": 549, "name": "_callTestMethod", "filename": 7}, {"line": 2918, "name": "wrapper", "filename": 3}, {"line": 10125, "name": "test_graph_break_compilation_metrics", "filename": 1}, {"line": 464, "name": "_fn", "filename": 8}, {"line": 10115, "name": "fn", "filename": 1}]}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.404000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "5922bc54db39a829d4160993cce4119c"} +V1206 15:20:13.926000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "bb6a3dc5077f86b72cd9a27f294c8718"} { - "name": "_compile.compile_inner", - "ts": 1723049492404415.8, - "args": null, + "name": "dynamo", + "ts": 1733527213926572.8, + "args": { + "compile_id": "0/0" + }, "ph": "B", + "cat": "dynamo_timed", + "tid": 0, "pid": 0 } -V0807 09:51:32.404000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "4e349dec4915b0fac8720db5f1d7b3f1"} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", 0]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/test/dynamo/test_misc.py", 1]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_dynamo/test_case.py", 2]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/testing/_internal/common_utils.py", 3]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/main.py", 4]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/runner.py", 5]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/suite.py", 6]} +V1206 15:20:13.928000 1543231 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/case.py", 7]} +V1206 15:20:13.928000 1543231 torch/_dynamo/convert_frame.py:961] {"dynamo_start": {"stack": [{"line": 12143, "name": "", "filename": 1}, {"line": 39, "name": "run_tests", "filename": 2}, {"line": 1353, "name": "run_tests", "filename": 3}, {"line": 102, "name": "__init__", "filename": 4}, {"line": 274, "name": "runTests", "filename": 4}, {"line": 217, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 678, "name": "__call__", "filename": 7}, {"line": 3234, "name": "run", "filename": 3}, {"line": 3206, "name": "_run_custom", "filename": 3}, {"line": 623, "name": "run", "filename": 7}, {"line": 579, "name": "_callTestMethod", "filename": 7}, {"line": 3099, "name": "wrapper", "filename": 3}, {"line": 10765, "name": "test_graph_break_compilation_metrics_on_failure", "filename": 1}, {"line": 10755, "name": "fn", "filename": 1}]}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:13.928000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2a27101ddf2d6616698cc3633c77fc3f"} { "name": "entire_frame_compile", - "ts": 1723049492404521.2, - "args": null, + "ts": 1733527213928251.2, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" + }, "ph": "B", + "cat": "dynamo_timed", + "tid": 0, "pid": 0 } -V0807 09:51:32.407000 2824488 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 0, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.407000 2824488 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 2, "dtype": "torch.float32", "device": "device(type='cpu')", "size": [4, 4], "is_leaf": true, "stride": [4, 1], "storage": 0, "view_func": "", "describer_id": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.407000 2824488 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 0, "id": 0, "source": "L['x']"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.411000 2824488 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 1, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 1} -V0807 09:51:32.411000 2824488 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 2, "dtype": "torch.float32", "device": "device(type='cpu')", "size": [4, 4], "is_leaf": true, "stride": [4, 1], "storage": 0, "view_func": "", "describer_id": 1}, "frame_id": 0, "frame_compile_id": 0, "attempt": 1} -V0807 09:51:32.411000 2824488 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 1, "id": 0, "source": "L['x']"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 1} -V0807 09:51:32.414000 2824488 torch/_dynamo/output_graph.py:1337] {"dynamo_output_graph": {"sizes": {"l_x_": [4, 4], "cos": [4, 4]}}, "frame_id": 0, "frame_compile_id": 0, "attempt": 1, "has_payload": "14f2dba4f0c5f60fdcbcde755d248243"} +V1206 15:20:13.931000 1543231 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 0, "describer_id": 0, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:13.931000 1543231 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 0, "ndim": 2, "dtype": "torch.float32", "device": "device(type='cpu')", "size": [4, 4], "is_leaf": true, "stride": [4, 1], "storage": 0, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:13.931000 1543231 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 0, "source": "L['x']"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:14.084000 1543231 torch/_dynamo/output_graph.py:1336] {"dynamo_output_graph": {"sizes": {"l_x_": [4, 4], "sin": [4, 4]}}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "289c7de0221d9f71a5513269eb25c1a2"} class GraphModule(torch.nn.Module): def forward(self, L_x_: "f32[4, 4][4, 1]cpu"): l_x_ = L_x_ - # File: /data/users/jjwu/a/pytorch/test/dynamo/test_misc.py:10116 in fn, code: x.cos() - cos: "f32[4, 4][4, 1]cpu" = l_x_.cos(); l_x_ = cos = None - return () + # File: /data/users/xmfan/a/pytorch/test/dynamo/test_misc.py:10756 in fn, code: return x.sin() + sin: "f32[4, 4][4, 1]cpu" = l_x_.sin(); l_x_ = None + return (sin,) -V0807 09:51:32.414000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "fb2fb938fb67faa5c285257ae0f30d1a"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1723049492414336.0, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:32.414000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "c23902719ca0aeb35cddd4f30e15b72b"} +V1206 15:20:14.084000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "7f609f2caa4686da5214135d7cf5c6fc"} { "name": "backend_compile", - "ts": 1723049492414417.8, - "args": null, + "ts": 1733527214084600.2, + "args": { + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" + }, "ph": "B", + "cat": "dynamo_timed", + "tid": 0, "pid": 0 } -V0807 09:51:32.414000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "1b13e4287eb465f804ccab5afeb981bd"} +V1206 15:20:14.085000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "09b2c72fc692bf5a0d9cac524e3ba68b"} { "name": "backend_compile", - "ts": 1723049492414569.2, - "args": null, - "ph": "E", - "pid": 0 - } -V0807 09:51:32.414000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "41ad13b634777f0eceb1985582be6e08"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1723049492414628.0, - "args": null, + "ts": 1733527214085078.5, + "args": { + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" + }, "ph": "E", + "cat": "dynamo_timed", + "tid": 0, "pid": 0 } -V0807 09:51:32.417000 2824488 torch/_dynamo/guards.py:2211] {"dynamo_cpp_guards_str": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 1, "has_payload": "be0edc824f8c9943106a5a56f84d3f98"} - - TREE_GUARD_MANAGER: - +- RootGuardManager - | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:462 in init_ambient_guards - | +- GLOBAL_STATE: ___check_global_state() - | +- GuardManager: source=L['x'], accessed_by=DictGetItemGuardAccessor(x) - | | +- TENSOR_MATCH: check_tensor(L['x'], Tensor, DispatchKeySet(CPU, BackendSelect, ADInplaceOrView, AutogradCPU), torch.float32, device=None, requires_grad=False, size=[4, 4], stride=[4, 1]) - | | +- NO_HASATTR: hasattr(L['x'], '_dynamo_dynamic_indices') == False - | +- GuardManager: source=G, accessed_by=GlobalsGuardAccessor - | | +- GuardManager: source=G['torch'], accessed_by=DictGetItemGuardAccessor(torch) - | | | +- ID_MATCH: ___check_obj_id(G['torch'], 140666006061280) - | | | +- GuardManager: source=G['torch']._dynamo, accessed_by=GetAttrGuardAccessor(_dynamo) - | | | | +- ID_MATCH: ___check_obj_id(G['torch']._dynamo, 140665945390464) - | | | | +- GuardManager: source=G['torch']._dynamo.graph_break, accessed_by=GetAttrGuardAccessor(graph_break) - | | | | | +- ID_MATCH: ___check_obj_id(G['torch']._dynamo.graph_break, 140664823287232) - -V0807 09:51:32.417000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "c42e7b412bcfec845489dbba83486d48"} +V1206 15:20:14.085000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "8eb8723e6aafa6856981ca893f6bc0c9"} { "name": "entire_frame_compile", - "ts": 1723049492417530.0, - "args": null, - "ph": "E", - "pid": 0 - } -V0807 09:51:32.417000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "17bf1e5f6d80155117b2cf71e3bcb29a"} - { - "name": "_compile.compile_inner", - "ts": 1723049492417603.2, - "args": null, - "ph": "E", - "pid": 0 - } -V0807 09:51:32.417000 2824488 torch/_dynamo/utils.py:780] {"compilation_metrics": {"compile_id": "0/0", "frame_key": "1", "co_name": "fn", "co_filename": "/data/users/jjwu/a/pytorch/test/dynamo/test_misc.py", "co_firstlineno": 10115, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 9, "shape_env_guard_count": 0, "graph_op_count": 1, "graph_node_count": 3, "graph_input_count": 1, "start_time": 1723049492.404317, "entire_frame_compile_time_s": 0.013243675231933594, "backend_compile_time_s": 0.00034117698669433594, "inductor_compile_time_s": null, "code_gen_time_s": null, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": ["'skip function graph_break in file /data/users/jjwu/a/pytorch/torch/_dynamo/decorators.py'"], "dynamo_time_before_restart_s": 0.006322145462036133, "has_guarded_code": true, "possibly_missed_reinplacing_opportunities": 0}, "frame_id": 0, "frame_compile_id": 0, "attempt": 1} -V0807 09:51:32.418000 2824488 torch/_dynamo/convert_frame.py:868] {"dynamo_start": {"stack": [{"line": 11137, "name": "", "filename": 1}, {"line": 38, "name": "run_tests", "filename": 2}, {"line": 1270, "name": "run_tests", "filename": 3}, {"line": 101, "name": "__init__", "filename": 4}, {"line": 271, "name": "runTests", "filename": 4}, {"line": 184, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 650, "name": "__call__", "filename": 7}, {"line": 3051, "name": "run", "filename": 3}, {"line": 3023, "name": "_run_custom", "filename": 3}, {"line": 591, "name": "run", "filename": 7}, {"line": 549, "name": "_callTestMethod", "filename": 7}, {"line": 2918, "name": "wrapper", "filename": 3}, {"line": 10125, "name": "test_graph_break_compilation_metrics", "filename": 1}, {"line": 464, "name": "_fn", "filename": 8}, {"line": 10117, "name": "fn", "filename": 1}, {"line": 10117, "name": "torch_dynamo_resume_in_fn_at_10117", "filename": 1}]}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.418000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "62a9731ed19442d613f1d0b530af14eb"} - { - "name": "_compile.compile_inner", - "ts": 1723049492418357.2, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:32.418000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "30545a1bf7efef4ca9cca35faa1ec3de"} - { - "name": "entire_frame_compile", - "ts": 1723049492418427.5, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:32.419000 2824488 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 3, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.419000 2824488 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 2, "dtype": "torch.float32", "device": "device(type='cpu')", "size": [4, 4], "is_leaf": true, "stride": [4, 1], "storage": 0, "view_func": "", "describer_id": 3}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.419000 2824488 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 3, "id": 0, "source": "L['x']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.421000 2824488 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 4, "size": 64}, "frame_id": 1, "frame_compile_id": 0, "attempt": 1} -V0807 09:51:32.421000 2824488 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 2, "dtype": "torch.float32", "device": "device(type='cpu')", "size": [4, 4], "is_leaf": true, "stride": [4, 1], "storage": 0, "view_func": "", "describer_id": 4}, "frame_id": 1, "frame_compile_id": 0, "attempt": 1} -V0807 09:51:32.421000 2824488 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 4, "id": 0, "source": "L['x']"}, "frame_id": 1, "frame_compile_id": 0, "attempt": 1} -V0807 09:51:32.423000 2824488 torch/_dynamo/output_graph.py:1337] {"dynamo_output_graph": {"sizes": {"l_x_": [4, 4], "sin": [4, 4]}}, "frame_id": 1, "frame_compile_id": 0, "attempt": 1, "has_payload": "982e9a7c38364664154acd1d0ebefaa3"} - class GraphModule(torch.nn.Module): - def forward(self, L_x_: "f32[4, 4][4, 1]cpu"): - l_x_ = L_x_ - - # File: /data/users/jjwu/a/pytorch/test/dynamo/test_misc.py:10118 in torch_dynamo_resume_in_fn_at_10117, code: x.sin() - sin: "f32[4, 4][4, 1]cpu" = l_x_.sin(); l_x_ = sin = None - return () - -V0807 09:51:32.423000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "1d72897e58f60785b16f3d20869623be"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1723049492423506.8, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:32.423000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "9b9d9a2486f8b7e8d0c40b3f383d3f09"} - { - "name": "backend_compile", - "ts": 1723049492423693.5, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:32.423000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "150523ccbde1f975220d1a2838bcd170"} - { - "name": "backend_compile", - "ts": 1723049492423789.2, - "args": null, + "ts": 1733527214085499.0, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" + }, "ph": "E", - "pid": 0 - } -V0807 09:51:32.423000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "7274bfa940bafe3a302f54894604e3ff"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1723049492423846.0, - "args": null, - "ph": "E", - "pid": 0 - } -V0807 09:51:32.426000 2824488 torch/_dynamo/guards.py:2211] {"dynamo_cpp_guards_str": {}, "frame_id": 1, "frame_compile_id": 0, "attempt": 1, "has_payload": "be0edc824f8c9943106a5a56f84d3f98"} - - TREE_GUARD_MANAGER: - +- RootGuardManager - | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:462 in init_ambient_guards - | +- GLOBAL_STATE: ___check_global_state() - | +- GuardManager: source=L['x'], accessed_by=DictGetItemGuardAccessor(x) - | | +- TENSOR_MATCH: check_tensor(L['x'], Tensor, DispatchKeySet(CPU, BackendSelect, ADInplaceOrView, AutogradCPU), torch.float32, device=None, requires_grad=False, size=[4, 4], stride=[4, 1]) - | | +- NO_HASATTR: hasattr(L['x'], '_dynamo_dynamic_indices') == False - | +- GuardManager: source=G, accessed_by=GlobalsGuardAccessor - | | +- GuardManager: source=G['torch'], accessed_by=DictGetItemGuardAccessor(torch) - | | | +- ID_MATCH: ___check_obj_id(G['torch'], 140666006061280) - | | | +- GuardManager: source=G['torch']._dynamo, accessed_by=GetAttrGuardAccessor(_dynamo) - | | | | +- ID_MATCH: ___check_obj_id(G['torch']._dynamo, 140665945390464) - | | | | +- GuardManager: source=G['torch']._dynamo.graph_break, accessed_by=GetAttrGuardAccessor(graph_break) - | | | | | +- ID_MATCH: ___check_obj_id(G['torch']._dynamo.graph_break, 140664823287232) + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:20:14.086000 1543231 torch/_dynamo/convert_frame.py:1011] {"artifact": {"name": "dynamo_error", "encoding": "string"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "8240b625682b3ef80eddf3423a91511c"} + Traceback (most recent call last): + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 989, in _compile + guarded_code = compile_inner(code, one_graph, hooks, transform) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 718, in compile_inner + return _compile_inner(code, one_graph, hooks, transform) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_utils_internal.py", line 95, in wrapper_function + return function(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 753, in _compile_inner + out_code = transform_code_object(code, transform) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/bytecode_transformation.py", line 1361, in transform_code_object + transformations(instructions, code_options) + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 231, in _fn + return fn(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 665, in transform + tracer.run() + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 2864, in run + super().run() + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 1053, in run + while self.step(): + ^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 963, in step + self.dispatch_table[inst.opcode](self, inst) + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 3044, in RETURN_VALUE + self._return(inst) + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 3029, in _return + self.output.compile_subgraph( + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1085, in compile_subgraph + self.compile_and_call_fx_graph(tx, list(reversed(stack_values)), root) + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1359, in compile_and_call_fx_graph + compiled_fn = self.call_user_compiler(gm) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1409, in call_user_compiler + return self._call_user_compiler(gm) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1460, in _call_user_compiler + raise BackendCompilerFailed(self.compiler_fn, e).with_traceback( + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1439, in _call_user_compiler + compiled_fn = compiler_fn(gm, self.example_inputs()) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/repro/after_dynamo.py", line 130, in __call__ + compiled_gm = compiler_fn(gm, example_inputs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/__init__.py", line 2353, in __call__ + return self.compiler_fn(model_, inputs_, **self.kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/test/dynamo/test_misc.py", line 10759, in broken_backend + raise RuntimeError("broken backend") + torch._dynamo.exc.BackendCompilerFailed: backend='broken_backend' raised: + RuntimeError: broken backend -V0807 09:51:32.426000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "49ab36075cbca8afadbe97390769ba16"} - { - "name": "entire_frame_compile", - "ts": 1723049492426141.2, - "args": null, - "ph": "E", - "pid": 0 - } -V0807 09:51:32.426000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "470360a71e2e7dc5ef2766ca058cb165"} - { - "name": "_compile.compile_inner", - "ts": 1723049492426211.0, - "args": null, - "ph": "E", - "pid": 0 - } -V0807 09:51:32.426000 2824488 torch/_dynamo/utils.py:780] {"compilation_metrics": {"compile_id": "1/0", "frame_key": "2", "co_name": "torch_dynamo_resume_in_fn_at_10117", "co_filename": "/data/users/jjwu/a/pytorch/test/dynamo/test_misc.py", "co_firstlineno": 10117, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 9, "shape_env_guard_count": 0, "graph_op_count": 1, "graph_node_count": 3, "graph_input_count": 1, "start_time": 1723049492.4183362, "entire_frame_compile_time_s": 0.007906198501586914, "backend_compile_time_s": 0.0003859996795654297, "inductor_compile_time_s": null, "code_gen_time_s": null, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": ["'skip function graph_break in file /data/users/jjwu/a/pytorch/torch/_dynamo/decorators.py'"], "dynamo_time_before_restart_s": 0.0023925304412841797, "has_guarded_code": true, "possibly_missed_reinplacing_opportunities": 0}, "frame_id": 1, "frame_compile_id": 0, "attempt": 1} -V0807 09:51:32.426000 2824488 torch/_dynamo/convert_frame.py:868] {"dynamo_start": {"stack": [{"line": 11137, "name": "", "filename": 1}, {"line": 38, "name": "run_tests", "filename": 2}, {"line": 1270, "name": "run_tests", "filename": 3}, {"line": 101, "name": "__init__", "filename": 4}, {"line": 271, "name": "runTests", "filename": 4}, {"line": 184, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 650, "name": "__call__", "filename": 7}, {"line": 3051, "name": "run", "filename": 3}, {"line": 3023, "name": "_run_custom", "filename": 3}, {"line": 591, "name": "run", "filename": 7}, {"line": 549, "name": "_callTestMethod", "filename": 7}, {"line": 2918, "name": "wrapper", "filename": 3}, {"line": 10125, "name": "test_graph_break_compilation_metrics", "filename": 1}, {"line": 464, "name": "_fn", "filename": 8}, {"line": 10117, "name": "fn", "filename": 1}, {"line": 10119, "name": "torch_dynamo_resume_in_fn_at_10117", "filename": 1}, {"line": 10119, "name": "torch_dynamo_resume_in_fn_at_10119", "filename": 1}]}, "frame_id": 2, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.426000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "cd67069b314de7443588126f4bb9caab"} - { - "name": "_compile.compile_inner", - "ts": 1723049492426820.8, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:32.426000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "535fe893ff4610c74adc1a5f325f9cdf"} - { - "name": "entire_frame_compile", - "ts": 1723049492426888.0, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:32.427000 2824488 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 6, "size": 64}, "frame_id": 2, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.427000 2824488 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 2, "dtype": "torch.float32", "device": "device(type='cpu')", "size": [4, 4], "is_leaf": true, "stride": [4, 1], "storage": 0, "view_func": "", "describer_id": 6}, "frame_id": 2, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.427000 2824488 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 6, "id": 0, "source": "L['x']"}, "frame_id": 2, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:32.429000 2824488 torch/_dynamo/output_graph.py:1337] {"dynamo_output_graph": {"sizes": {"l_x_": [4, 4], "cos": [4, 4]}}, "frame_id": 2, "frame_compile_id": 0, "attempt": 0, "has_payload": "a6969283b355d59fdf903f7c44eafd31"} - class GraphModule(torch.nn.Module): - def forward(self, L_x_: "f32[4, 4][4, 1]cpu"): - l_x_ = L_x_ - - # File: /data/users/jjwu/a/pytorch/test/dynamo/test_misc.py:10120 in torch_dynamo_resume_in_fn_at_10119, code: return x.cos() - cos: "f32[4, 4][4, 1]cpu" = l_x_.cos(); l_x_ = None - return (cos,) - -V0807 09:51:32.429000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "10ccd8fe9d672960bbfdd6b77179f982"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1723049492429474.8, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:32.429000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "d3f5288a1b1dbe58fc864196ab3d473d"} - { - "name": "backend_compile", - "ts": 1723049492429555.2, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:32.429000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "c8167a747bfced74d17f4774d8c153e0"} - { - "name": "backend_compile", - "ts": 1723049492429640.0, - "args": null, - "ph": "E", - "pid": 0 - } -V0807 09:51:32.429000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "93de54b9806ab039e3f47a67ed0ba30d"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1723049492429692.5, - "args": null, - "ph": "E", - "pid": 0 - } -V0807 09:51:32.430000 2824488 torch/_dynamo/guards.py:2211] {"dynamo_cpp_guards_str": {}, "frame_id": 2, "frame_compile_id": 0, "attempt": 0, "has_payload": "85f4665e3933f3363f5cc2aee88cf803"} + Set TORCH_LOGS="+dynamo" and TORCHDYNAMO_VERBOSE=1 for more information - TREE_GUARD_MANAGER: - +- RootGuardManager - | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:462 in init_ambient_guards - | +- GLOBAL_STATE: ___check_global_state() - | +- GuardManager: source=L['x'], accessed_by=DictGetItemGuardAccessor(x) - | | +- TENSOR_MATCH: check_tensor(L['x'], Tensor, DispatchKeySet(CPU, BackendSelect, ADInplaceOrView, AutogradCPU), torch.float32, device=None, requires_grad=False, size=[4, 4], stride=[4, 1]) - | | +- NO_HASATTR: hasattr(L['x'], '_dynamo_dynamic_indices') == False -V0807 09:51:32.430000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "18b8a657af01bb59c3d74b7ef8ff9203"} - { - "name": "entire_frame_compile", - "ts": 1723049492430901.5, - "args": null, - "ph": "E", - "pid": 0 - } -V0807 09:51:32.430000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "1b512edd840f7ea466b0b8730253cd1f"} - { - "name": "_compile.compile_inner", - "ts": 1723049492430971.0, - "args": null, +V1206 15:20:14.089000 1543231 torch/_dynamo/utils.py:1045] {"compilation_metrics": {"aot_autograd_cumulative_compile_time_us": 478, "dynamo_cumulative_compile_time_us": 157247, "frame_key": "1", "co_name": "fn", "co_filename": "/data/users/xmfan/a/pytorch/test/dynamo/test_misc.py", "co_firstlineno": 10755, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": null, "shape_env_guard_count": null, "graph_op_count": null, "graph_node_count": null, "graph_input_count": null, "fail_type": "BackendCompilerFailed", "fail_reason": "backend='broken_backend' raised:\nRuntimeError: broken backend", "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.158574662, "has_guarded_code": false, "config_suppress_errors": true, "config_inline_inbuilt_nn_modules": true, "specialize_float": false, "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": true, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._prims\", \"torch.testing\", \"torch.distributions\", \"torch._refs\", \"torch._decomp\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": false, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}", "is_forward": true, "dynamo_compile_time_before_restart_us": 158574}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:14.089000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "99d536cc3a0e4bcc211319f5a4656991"} + { + "name": "dynamo", + "ts": 1733527214089673.0, + "args": { + "compile_id": "0/0", + "frame_key": "1", + "co_name": "fn", + "co_filename": "/data/users/xmfan/a/pytorch/test/dynamo/test_misc.py", + "co_firstlineno": 10755, + "cache_size": 0, + "accumulated_cache_size": 0, + "guard_count": null, + "shape_env_guard_count": null, + "graph_op_count": null, + "graph_node_count": null, + "graph_input_count": null, + "fail_type": "BackendCompilerFailed", + "fail_reason": "backend='broken_backend' raised:\nRuntimeError: broken backend", + "fail_user_frame_filename": null, + "fail_user_frame_lineno": null, + "non_compliant_ops": [], + "compliant_custom_ops": [], + "restart_reasons": [], + "dynamo_time_before_restart_s": 0.158574662, + "has_guarded_code": false, + "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": true, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._prims\", \"torch.testing\", \"torch.distributions\", \"torch._refs\", \"torch._decomp\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": false, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}" + }, "ph": "E", + "cat": "dynamo_timed", + "tid": 0, "pid": 0 - } -V0807 09:51:32.431000 2824488 torch/_dynamo/utils.py:780] {"compilation_metrics": {"compile_id": "2/0", "frame_key": "3", "co_name": "torch_dynamo_resume_in_fn_at_10119", "co_filename": "/data/users/jjwu/a/pytorch/test/dynamo/test_misc.py", "co_firstlineno": 10119, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 6, "shape_env_guard_count": 0, "graph_op_count": 1, "graph_node_count": 3, "graph_input_count": 1, "start_time": 1723049492.4268005, "entire_frame_compile_time_s": 0.0042018890380859375, "backend_compile_time_s": 0.00026297569274902344, "inductor_compile_time_s": null, "code_gen_time_s": null, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "possibly_missed_reinplacing_opportunities": 0}, "frame_id": 2, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:33.443000 2824488 torch/_dynamo/convert_frame.py:868] {"dynamo_start": {"stack": [{"line": 11137, "name": "", "filename": 1}, {"line": 38, "name": "run_tests", "filename": 2}, {"line": 1270, "name": "run_tests", "filename": 3}, {"line": 101, "name": "__init__", "filename": 4}, {"line": 271, "name": "runTests", "filename": 4}, {"line": 184, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 650, "name": "__call__", "filename": 7}, {"line": 3051, "name": "run", "filename": 3}, {"line": 3023, "name": "_run_custom", "filename": 3}, {"line": 591, "name": "run", "filename": 7}, {"line": 549, "name": "_callTestMethod", "filename": 7}, {"line": 2918, "name": "wrapper", "filename": 3}, {"line": 10165, "name": "test_graph_break_compilation_metrics_on_failure", "filename": 1}, {"line": 464, "name": "_fn", "filename": 8}, {"line": 10155, "name": "fn", "filename": 1}]}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:33.443000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "eef75bcdb35b95b75e2f35fa877d41a9"} - { - "name": "_compile.compile_inner", - "ts": 1723049493443444.5, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:33.443000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "354889097421a515ce496163d74a9fa3"} - { - "name": "entire_frame_compile", - "ts": 1723049493443541.5, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:33.444000 2824488 torch/_subclasses/meta_utils.py:204] {"describe_storage": {"id": 0, "describer_id": 8, "size": 64}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:33.444000 2824488 torch/_subclasses/meta_utils.py:417] {"describe_tensor": {"id": 0, "ndim": 2, "dtype": "torch.float32", "device": "device(type='cpu')", "size": [4, 4], "is_leaf": true, "stride": [4, 1], "storage": 0, "view_func": "", "describer_id": 8}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:33.444000 2824488 torch/_subclasses/meta_utils.py:1640] {"describe_source": {"describer_id": 8, "id": 0, "source": "L['x']"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0807 09:51:33.446000 2824488 torch/_dynamo/output_graph.py:1337] {"dynamo_output_graph": {"sizes": {"l_x_": [4, 4], "sin": [4, 4]}}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "ba845d3191efd77996f11649d13de94e"} - class GraphModule(torch.nn.Module): - def forward(self, L_x_: "f32[4, 4][4, 1]cpu"): - l_x_ = L_x_ - - # File: /data/users/jjwu/a/pytorch/test/dynamo/test_misc.py:10156 in fn, code: return x.sin() - sin: "f32[4, 4][4, 1]cpu" = l_x_.sin(); l_x_ = None - return (sin,) - -V0807 09:51:33.446000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "8615ceb16d467000ba450ddee01f45d1"} - { - "name": "OutputGraph.call_user_compiler", - "ts": 1723049493446954.2, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:33.447000 2824488 torch/_dynamo/utils.py:871] {"chromium_event": {}, "has_payload": "c461c120aa4c7e0f92d09ef62d92a5d1"} - { - "name": "backend_compile", - "ts": 1723049493447036.8, - "args": null, - "ph": "B", - "pid": 0 - } -V0807 09:51:33.447000 2824488 torch/_dynamo/utils.py:780] {"compilation_metrics": {"compile_id": "0/0", "frame_key": "1", "co_name": "fn", "co_filename": "/data/users/jjwu/a/pytorch/test/dynamo/test_misc.py", "co_firstlineno": 10155, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": null, "shape_env_guard_count": null, "graph_op_count": null, "graph_node_count": null, "graph_input_count": null, "start_time": 1723049493.4433863, "entire_frame_compile_time_s": null, "backend_compile_time_s": null, "inductor_compile_time_s": null, "code_gen_time_s": null, "fail_type": "", "fail_reason": "backend='broken_backend' raised:\nRuntimeError: broken backend", "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0039255619049072266, "has_guarded_code": false, "possibly_missed_reinplacing_opportunities": null}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} + } \ No newline at end of file diff --git a/tests/inputs/comp_failure.log b/tests/inputs/comp_failure.log index c8ea39b..27d67d8 100644 --- a/tests/inputs/comp_failure.log +++ b/tests/inputs/comp_failure.log @@ -1,19 +1,178 @@ -V0404 09:04:28.542000 140560297141248 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/test/dynamo/test_misc.py", 0]} -V0404 09:04:28.543000 140560297141248 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/test_case.py", 1]} -V0404 09:04:28.543000 140560297141248 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/testing/_internal/common_utils.py", 2]} -V0404 09:04:28.543000 140560297141248 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/main.py", 3]} -V0404 09:04:28.543000 140560297141248 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/runner.py", 4]} -V0404 09:04:28.543000 140560297141248 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/suite.py", 5]} -V0404 09:04:28.543000 140560297141248 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/case.py", 6]} -V0404 09:04:28.543000 140560297141248 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/eval_frame.py", 7]} -V0404 09:04:28.543000 140560297141248 torch/_dynamo/convert_frame.py:672] {"dynamo_start": {"stack": [{"line": 10079, "name": "", "filename": 0}, {"line": 41, "name": "run_tests", "filename": 1}, {"line": 1167, "name": "run_tests", "filename": 2}, {"line": 101, "name": "__init__", "filename": 3}, {"line": 271, "name": "runTests", "filename": 3}, {"line": 184, "name": "run", "filename": 4}, {"line": 84, "name": "__call__", "filename": 5}, {"line": 122, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 5}, {"line": 122, "name": "run", "filename": 5}, {"line": 650, "name": "__call__", "filename": 6}, {"line": 2868, "name": "run", "filename": 2}, {"line": 2840, "name": "_run_custom", "filename": 2}, {"line": 591, "name": "run", "filename": 6}, {"line": 549, "name": "_callTestMethod", "filename": 6}, {"line": 2741, "name": "wrapper", "filename": 2}, {"line": 9599, "name": "test_graph_break_compilation_metrics_on_failure", "filename": 0}, {"line": 410, "name": "_fn", "filename": 7}]}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0404 09:04:28.552000 140560297141248 torch/_dynamo/output_graph.py:1139] {"dynamo_output_graph": {"sizes": {"l_x_": [4, 4], "sin": [4, 4]}}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "c42a102d057a4af133785f9775eefaef"} +V1206 15:20:13.926000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "bb6a3dc5077f86b72cd9a27f294c8718"} + { + "name": "dynamo", + "ts": 1733527213926572.8, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", 0]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/test/dynamo/test_misc.py", 1]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_dynamo/test_case.py", 2]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/testing/_internal/common_utils.py", 3]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/main.py", 4]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/runner.py", 5]} +V1206 15:20:13.927000 1543231 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/suite.py", 6]} +V1206 15:20:13.928000 1543231 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/case.py", 7]} +V1206 15:20:13.928000 1543231 torch/_dynamo/convert_frame.py:961] {"dynamo_start": {"stack": [{"line": 12143, "name": "", "filename": 1}, {"line": 39, "name": "run_tests", "filename": 2}, {"line": 1353, "name": "run_tests", "filename": 3}, {"line": 102, "name": "__init__", "filename": 4}, {"line": 274, "name": "runTests", "filename": 4}, {"line": 217, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 678, "name": "__call__", "filename": 7}, {"line": 3234, "name": "run", "filename": 3}, {"line": 3206, "name": "_run_custom", "filename": 3}, {"line": 623, "name": "run", "filename": 7}, {"line": 579, "name": "_callTestMethod", "filename": 7}, {"line": 3099, "name": "wrapper", "filename": 3}, {"line": 10765, "name": "test_graph_break_compilation_metrics_on_failure", "filename": 1}, {"line": 10755, "name": "fn", "filename": 1}]}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:13.928000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2a27101ddf2d6616698cc3633c77fc3f"} + { + "name": "entire_frame_compile", + "ts": 1733527213928251.2, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:20:13.931000 1543231 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 0, "describer_id": 0, "size": 64}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:13.931000 1543231 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 0, "ndim": 2, "dtype": "torch.float32", "device": "device(type='cpu')", "size": [4, 4], "is_leaf": true, "stride": [4, 1], "storage": 0, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:13.931000 1543231 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 0, "source": "L['x']"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:14.084000 1543231 torch/_dynamo/output_graph.py:1336] {"dynamo_output_graph": {"sizes": {"l_x_": [4, 4], "sin": [4, 4]}}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "289c7de0221d9f71a5513269eb25c1a2"} class GraphModule(torch.nn.Module): - def forward(self, L_x_ : torch.Tensor): + def forward(self, L_x_: "f32[4, 4][4, 1]cpu"): l_x_ = L_x_ - # File: /data/users/jjwu/a/pytorch/test/dynamo/test_misc.py:9590 in fn, code: return x.sin() - sin = l_x_.sin(); l_x_ = None + # File: /data/users/xmfan/a/pytorch/test/dynamo/test_misc.py:10756 in fn, code: return x.sin() + sin: "f32[4, 4][4, 1]cpu" = l_x_.sin(); l_x_ = None return (sin,) -V0404 09:04:28.553000 140560297141248 torch/_dynamo/utils.py:685] {"compilation_metrics": {"frame_key": "1", "co_name": "fn", "co_filename": "/data/users/jjwu/a/pytorch/test/dynamo/test_misc.py", "co_firstlineno": 9589, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": null, "shape_env_guard_count": null, "graph_op_count": null, "graph_node_count": null, "graph_input_count": null, "start_time": 1712246668.5434601, "entire_frame_compile_time_s": null, "backend_compile_time_s": null, "inductor_compile_time_s": null, "code_gen_time_s": null, "fail_type": "", "fail_reason": "backend='broken_backend' raised:\nRuntimeError: broken backend", "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.009927034378051758}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:14.084000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "7f609f2caa4686da5214135d7cf5c6fc"} + { + "name": "backend_compile", + "ts": 1733527214084600.2, + "args": { + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:20:14.085000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "09b2c72fc692bf5a0d9cac524e3ba68b"} + { + "name": "backend_compile", + "ts": 1733527214085078.5, + "args": { + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:20:14.085000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "8eb8723e6aafa6856981ca893f6bc0c9"} + { + "name": "entire_frame_compile", + "ts": 1733527214085499.0, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:20:14.086000 1543231 torch/_dynamo/convert_frame.py:1011] {"artifact": {"name": "dynamo_error", "encoding": "string"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "8240b625682b3ef80eddf3423a91511c"} + Traceback (most recent call last): + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 989, in _compile + guarded_code = compile_inner(code, one_graph, hooks, transform) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 718, in compile_inner + return _compile_inner(code, one_graph, hooks, transform) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_utils_internal.py", line 95, in wrapper_function + return function(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 753, in _compile_inner + out_code = transform_code_object(code, transform) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/bytecode_transformation.py", line 1361, in transform_code_object + transformations(instructions, code_options) + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 231, in _fn + return fn(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", line 665, in transform + tracer.run() + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 2864, in run + super().run() + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 1053, in run + while self.step(): + ^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 963, in step + self.dispatch_table[inst.opcode](self, inst) + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 3044, in RETURN_VALUE + self._return(inst) + File "/data/users/xmfan/a/pytorch/torch/_dynamo/symbolic_convert.py", line 3029, in _return + self.output.compile_subgraph( + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1085, in compile_subgraph + self.compile_and_call_fx_graph(tx, list(reversed(stack_values)), root) + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1359, in compile_and_call_fx_graph + compiled_fn = self.call_user_compiler(gm) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1409, in call_user_compiler + return self._call_user_compiler(gm) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1460, in _call_user_compiler + raise BackendCompilerFailed(self.compiler_fn, e).with_traceback( + File "/data/users/xmfan/a/pytorch/torch/_dynamo/output_graph.py", line 1439, in _call_user_compiler + compiled_fn = compiler_fn(gm, self.example_inputs()) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/_dynamo/repro/after_dynamo.py", line 130, in __call__ + compiled_gm = compiler_fn(gm, example_inputs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/torch/__init__.py", line 2353, in __call__ + return self.compiler_fn(model_, inputs_, **self.kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/data/users/xmfan/a/pytorch/test/dynamo/test_misc.py", line 10759, in broken_backend + raise RuntimeError("broken backend") + torch._dynamo.exc.BackendCompilerFailed: backend='broken_backend' raised: + RuntimeError: broken backend + + Set TORCH_LOGS="+dynamo" and TORCHDYNAMO_VERBOSE=1 for more information + + +V1206 15:20:14.089000 1543231 torch/_dynamo/utils.py:1045] {"compilation_metrics": {"aot_autograd_cumulative_compile_time_us": 478, "dynamo_cumulative_compile_time_us": 157247, "frame_key": "1", "co_name": "fn", "co_filename": "/data/users/xmfan/a/pytorch/test/dynamo/test_misc.py", "co_firstlineno": 10755, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": null, "shape_env_guard_count": null, "graph_op_count": null, "graph_node_count": null, "graph_input_count": null, "fail_type": "BackendCompilerFailed", "fail_reason": "backend='broken_backend' raised:\nRuntimeError: broken backend", "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": [], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.158574662, "has_guarded_code": false, "config_suppress_errors": true, "config_inline_inbuilt_nn_modules": true, "specialize_float": false, "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": true, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._prims\", \"torch.testing\", \"torch.distributions\", \"torch._refs\", \"torch._decomp\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": false, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}", "is_forward": true, "dynamo_compile_time_before_restart_us": 158574}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:20:14.089000 1543231 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "99d536cc3a0e4bcc211319f5a4656991"} + { + "name": "dynamo", + "ts": 1733527214089673.0, + "args": { + "compile_id": "0/0", + "frame_key": "1", + "co_name": "fn", + "co_filename": "/data/users/xmfan/a/pytorch/test/dynamo/test_misc.py", + "co_firstlineno": 10755, + "cache_size": 0, + "accumulated_cache_size": 0, + "guard_count": null, + "shape_env_guard_count": null, + "graph_op_count": null, + "graph_node_count": null, + "graph_input_count": null, + "fail_type": "BackendCompilerFailed", + "fail_reason": "backend='broken_backend' raised:\nRuntimeError: broken backend", + "fail_user_frame_filename": null, + "fail_user_frame_lineno": null, + "non_compliant_ops": [], + "compliant_custom_ops": [], + "restart_reasons": [], + "dynamo_time_before_restart_s": 0.158574662, + "has_guarded_code": false, + "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": true, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._prims\", \"torch.testing\", \"torch.distributions\", \"torch._refs\", \"torch._decomp\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": false, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } \ No newline at end of file diff --git a/tests/inputs/simple.log b/tests/inputs/simple.log index dec360b..69b998a 100644 --- a/tests/inputs/simple.log +++ b/tests/inputs/simple.log @@ -1,80 +1,634 @@ -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py", 0]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_inductor/test_case.py", 1]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/test_case.py", 2]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/testing/_internal/common_utils.py", 3]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/main.py", 4]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/runner.py", 5]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/suite.py", 6]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/unittest/case.py", 7]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch-env/lib/python3.10/contextlib.py", 8]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/nn/modules/module.py", 9]} -V0401 08:54:21.881000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/eval_frame.py", 10]} -V0401 08:54:21.882000 140424060892160 torch/_dynamo/convert_frame.py:672] {"dynamo_start": {"stack": [{"line": 10031, "name": "", "filename": 0}, {"line": 14, "name": "run_tests", "filename": 1}, {"line": 41, "name": "run_tests", "filename": 2}, {"line": 1165, "name": "run_tests", "filename": 3}, {"line": 101, "name": "__init__", "filename": 4}, {"line": 271, "name": "runTests", "filename": 4}, {"line": 184, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 650, "name": "__call__", "filename": 7}, {"line": 2866, "name": "run", "filename": 3}, {"line": 2838, "name": "_run_custom", "filename": 3}, {"line": 591, "name": "run", "filename": 7}, {"line": 549, "name": "_callTestMethod", "filename": 7}, {"line": 2739, "name": "wrapper", "filename": 3}, {"line": 9214, "name": "new_test", "filename": 0}, {"line": 79, "name": "inner", "filename": 8}, {"line": 8845, "name": "test_custom_op_fixed_layout_channels_last", "filename": 0}, {"line": 1527, "name": "_wrapped_call_impl", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}, {"line": 450, "name": "_fn", "filename": 10}, {"line": 1527, "name": "_wrapped_call_impl", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}]}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0401 08:54:21.895000 140424060892160 torch/_dynamo/output_graph.py:1189] {"dynamo_output_graph": {"sizes": {"l_x_": [1, 320, 128, 128], "out": [1, 320, 128, 128], "out_2": [1, 320, 128, 128], "out_3": [1, 320, 128, 128]}}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "74cabbbada68afbad8a921c47aa2b317"} +V1206 15:18:15.925000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "has_payload": "490a3a38ce10390374e403dd6e063c7f"} + { + "name": "compile_file", + "ts": 1733527095925179.5, + "args": { + "compile_id": "None" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:16.845000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "has_payload": "6762d47fdbf80071626529f25dc69013"} + { + "name": "compile_file", + "ts": 1733527096845517.0, + "args": { + "compile_id": "None" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:17.965000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "has_payload": "acbf9ac4749cce84db0b5b9acd4b26e5"} + { + "name": "compile_file", + "ts": 1733527097965329.8, + "args": { + "compile_id": "None" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:18.815000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "has_payload": "6050df090b76feb2a53d6cc2fdb23908"} + { + "name": "compile_file", + "ts": 1733527098815455.0, + "args": { + "compile_id": "None" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:19.883000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "has_payload": "f35f8d0fcca57799f1a731e665c95987"} + { + "name": "compile_file", + "ts": 1733527099883504.5, + "args": { + "compile_id": "None" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:19.999000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "has_payload": "01af5c3df1e6a6aa1747dab60c80c0b9"} + { + "name": "compile_file", + "ts": 1733527099999552.0, + "args": { + "compile_id": "None" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:20.252000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2edd807b9a63e3b4c6d50b6297bb4677"} + { + "name": "dynamo", + "ts": 1733527100252271.5, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:20.253000 1500233 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_dynamo/convert_frame.py", 0]} +V1206 15:18:20.253000 1500233 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py", 1]} +V1206 15:18:20.253000 1500233 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_inductor/test_case.py", 2]} +V1206 15:18:20.253000 1500233 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_dynamo/test_case.py", 3]} +V1206 15:18:20.253000 1500233 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/testing/_internal/common_utils.py", 4]} +V1206 15:18:20.254000 1500233 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/main.py", 5]} +V1206 15:18:20.254000 1500233 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/runner.py", 6]} +V1206 15:18:20.254000 1500233 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/suite.py", 7]} +V1206 15:18:20.254000 1500233 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/unittest/case.py", 8]} +V1206 15:18:20.254000 1500233 torch/_logging/structured.py:22] {"str": ["/home/xmfan/local/a/pytorch-env/lib/python3.11/contextlib.py", 9]} +V1206 15:18:20.254000 1500233 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/nn/modules/module.py", 10]} +V1206 15:18:20.254000 1500233 torch/_logging/structured.py:22] {"str": ["/data/users/xmfan/a/pytorch/torch/_dynamo/eval_frame.py", 11]} +V1206 15:18:20.254000 1500233 torch/_dynamo/convert_frame.py:961] {"dynamo_start": {"stack": [{"line": 13037, "name": "", "filename": 1}, {"line": 15, "name": "run_tests", "filename": 2}, {"line": 39, "name": "run_tests", "filename": 3}, {"line": 1353, "name": "run_tests", "filename": 4}, {"line": 102, "name": "__init__", "filename": 5}, {"line": 274, "name": "runTests", "filename": 5}, {"line": 217, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 84, "name": "__call__", "filename": 7}, {"line": 122, "name": "run", "filename": 7}, {"line": 678, "name": "__call__", "filename": 8}, {"line": 3234, "name": "run", "filename": 4}, {"line": 3206, "name": "_run_custom", "filename": 4}, {"line": 623, "name": "run", "filename": 8}, {"line": 579, "name": "_callTestMethod", "filename": 8}, {"line": 3099, "name": "wrapper", "filename": 4}, {"line": 11906, "name": "new_test", "filename": 1}, {"line": 81, "name": "inner", "filename": 9}, {"line": 11160, "name": "test_custom_op_fixed_layout_channels_last", "filename": 1}, {"line": 1739, "name": "_wrapped_call_impl", "filename": 10}, {"line": 1750, "name": "_call_impl", "filename": 10}, {"line": 573, "name": "_fn", "filename": 11}, {"line": 1739, "name": "_wrapped_call_impl", "filename": 10}, {"line": 11123, "name": "forward", "filename": 1}]}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:18:20.254000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "0a2ce00ecd7d0ef78918976e3561193a"} + { + "name": "entire_frame_compile", + "ts": 1733527100254653.2, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:20.575000 1500233 torch/_subclasses/meta_utils.py:241] {"describe_storage": {"id": 0, "describer_id": 0, "size": 20971520}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:18:20.575000 1500233 torch/_subclasses/meta_utils.py:454] {"describe_tensor": {"id": 0, "ndim": 4, "dtype": "torch.float32", "device": "device(type='cuda', index=0)", "size": [1, 320, 128, 128], "is_leaf": true, "stride": [5242880, 1, 40960, 320], "storage": 0, "view_func": "", "describer_id": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:18:20.576000 1500233 torch/_subclasses/meta_utils.py:1779] {"describe_source": {"describer_id": 0, "id": 0, "source": "L['x']"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:18:20.599000 1500233 torch/_dynamo/output_graph.py:1336] {"dynamo_output_graph": {"sizes": {"l_x_": [1, 320, 128, 128], "out": [1, 320, 128, 128], "input_1": [1, 320, 128, 128], "out_1": [1, 320, 128, 128]}}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "43df19c7545f412eca248f900fce6ce5"} class GraphModule(torch.nn.Module): - def forward(self, L_x_ : torch.Tensor): + def forward(self, L_x_: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0"): l_x_ = L_x_ - # File: /data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py:8809 in helper, code: out = F.gelu(x) - out = torch._C._nn.gelu(l_x_); l_x_ = None + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11119 in helper, code: out = F.gelu(x) + out: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch._C._nn.gelu(l_x_); l_x_ = None - # File: /data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py:8810 in helper, code: out = self.in_layers(out) - out_2 = self.L__self___in_layers_0(out); out = None + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11120 in helper, code: out = self.in_layers(out) + input_1: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.nn.functional.dropout(out, 0.1, True, False); out = None - # File: /data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py:8815 in forward, code: out = torch.ops.test.baz(out) - out_3 = torch.ops.test.baz(out_2); out_2 = None - return (out_3,) + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11125 in forward, code: out = torch.ops.test.baz(out) + out_1: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.test.baz(input_1); input_1 = None + return (out_1,) -V0401 08:54:21.914000 140424060892160 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:112] {"aot_forward_graph": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "9d0885575d7f020cb3a1996185650901"} +V1206 15:18:20.600000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "b02af7e6b7baf7b21e5166764db0a66f"} + { + "name": "backend_compile", + "ts": 1733527100600090.5, + "args": { + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:20.600000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "632df8822c4c5df2079dfd71f5827dc0"} + { + "name": "_recursive_pre_grad_passes", + "ts": 1733527100600377.2, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:20.612000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "ca25b04b9513de1a0144b92d0d669023"} + { + "name": "_recursive_pre_grad_passes", + "ts": 1733527100612301.8, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:20.613000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "ed067c57dd8cd9be0b9084c7c54c3c2e"} + { + "name": "inductor_codecache_torch_key", + "ts": 1733527100613547.5, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:20.687000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "80c4f441db0b0d76a817bcb027333d90"} + { + "name": "inductor_codecache_torch_key", + "ts": 1733527100687165.2, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.086000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2ae002c6629696bbf0b79d8f1b7e5969"} + { + "name": "create_aot_dispatcher_function", + "ts": 1733527101086489.5, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.105000 1500233 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:214] {"artifact": {"name": "aot_forward_graph_fw_metadata", "encoding": "string"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "619a91cc6a2a9ea21ce0754dd5524cda"} + ViewAndMutationMeta(input_info=[InputAliasInfo(is_leaf=True, + mutates_data=False, + mutates_metadata=False, + mutations_hidden_from_autograd=True, + mutations_under_no_grad_or_inference_mode=False, + mutation_inductor_storage_resize=False, + mutates_storage_metadata=False, + requires_grad=False, + keep_input_mutations=True)], + output_info=[OutputAliasInfo(output_type=, + raw_type=, + base_idx=None, + dynamic_dims=set(), + requires_grad=False, + functional_tensor=None)], + num_intermediate_bases=0, + keep_input_mutations=True, + traced_tangents=[], + subclass_inp_meta=[PlainTensorMeta(unwrapped_idx=0, + memory_format=None)], + subclass_fw_graph_out_meta=[PlainTensorMeta(unwrapped_idx=0, + memory_format=None)], + subclass_tangent_meta=[], + is_train=False, + traced_tangent_metas=None, + num_symints_saved_for_bw=None, + grad_enabled_mutation=None, + deterministic=None, + static_input_indices=[], + tokens={}, + indices_of_inputs_that_requires_grad_with_mutations_in_bw=[], + bw_donated_idxs=None, + num_backward_tokens=0) +V1206 15:18:21.106000 1500233 torch/_functorch/_aot_autograd/dispatch_and_compile_graph.py:232] {"aot_inference_graph": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "9217735d824ab4b7c7cb21827ecb4673"} class (torch.nn.Module): - def forward(self, arg0_1: "f32[1, 320, 128, 128]"): - # File: /data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py:8809 in helper, code: out = F.gelu(x) - mul: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(arg0_1, 0.5) - mul_1: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(arg0_1, 0.7071067811865476); arg0_1 = None - erf: "f32[1, 320, 128, 128]" = torch.ops.aten.erf.default(mul_1); mul_1 = None - add: "f32[1, 320, 128, 128]" = torch.ops.aten.add.Tensor(erf, 1); erf = None - mul_2: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(mul, add); mul = add = None + def forward(self, arg0_1: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0"): + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11119 in helper, code: out = F.gelu(x) + mul: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(arg0_1, 0.5) + mul_1: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(arg0_1, 0.7071067811865476); arg0_1 = None + erf: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.erf.default(mul_1); mul_1 = None + add: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.add.Tensor(erf, 1); erf = None + mul_2: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(mul, add); mul = add = None - # File: /data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py:8810 in helper, code: out = self.in_layers(out) - rand: "f32[1, 320, 128, 128]" = torch.ops.aten.rand.default([1, 320, 128, 128], dtype = torch.float32, device = device(type='cuda', index=0), pin_memory = False) - convert_element_type: "f32[1, 320, 128, 128]" = torch.ops.prims.convert_element_type.default(rand, torch.float32); rand = None - clone: "f32[1, 320, 128, 128]" = torch.ops.aten.clone.default(convert_element_type, memory_format = torch.channels_last); convert_element_type = None - gt: "b8[1, 320, 128, 128]" = torch.ops.aten.gt.Scalar(clone, 0.1); clone = None - mul_3: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(gt, mul_2); gt = mul_2 = None - mul_4: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(mul_3, 1.1111111111111112); mul_3 = None + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11120 in helper, code: out = self.in_layers(out) + rand: "f32[1, 320, 128, 128][5242880, 16384, 128, 1]cuda:0" = torch.ops.aten.rand.default([1, 320, 128, 128], dtype = torch.float32, device = device(type='cuda', index=0), pin_memory = False) + convert_element_type: "f32[1, 320, 128, 128][5242880, 16384, 128, 1]cuda:0" = torch.ops.prims.convert_element_type.default(rand, torch.float32); rand = None + clone: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.clone.default(convert_element_type, memory_format = torch.channels_last); convert_element_type = None + gt: "b8[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.gt.Scalar(clone, 0.1); clone = None + mul_3: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(gt, mul_2); gt = mul_2 = None + mul_4: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(mul_3, 1.1111111111111112); mul_3 = None - # File: /data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py:8815 in forward, code: out = torch.ops.test.baz(out) - baz: "f32[1, 320, 128, 128]" = torch.ops.test.baz.default(mul_4); mul_4 = None + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11125 in forward, code: out = torch.ops.test.baz(out) + baz: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.test.baz.default(mul_4); mul_4 = None return (baz,) -V0401 08:54:22.062000 140424060892160 torch/_inductor/compile_fx.py:650] {"inductor_post_grad_graph": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "cf729e565df6880d92548bb46eff9ab5"} - class (torch.nn.Module): - def forward(self, arg0_1: "f32[1, 320, 128, 128]"): - # File: /data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py:8809 in helper, code: out = F.gelu(x) - mul: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(arg0_1, 0.5) - mul_1: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(arg0_1, 0.7071067811865476); arg0_1 = None - erf: "f32[1, 320, 128, 128]" = torch.ops.aten.erf.default(mul_1); mul_1 = None - add: "f32[1, 320, 128, 128]" = torch.ops.aten.add.Tensor(erf, 1); erf = None - mul_2: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(mul, add); mul = add = None +V1206 15:18:21.106000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "404b42df98884a4e0973f79434379f9b"} + { + "name": "compile_fx..fw_compiler_base", + "ts": 1733527101106855.5, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.107000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "ef2aa74274539117c0be992171939d5b"} + { + "name": "_recursive_joint_graph_passes", + "ts": 1733527101107196.5, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.196000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "a6e306d9d22bfce0b4cbd732937552d9"} + { + "name": "_recursive_joint_graph_passes", + "ts": 1733527101196659.2, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.197000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "6c62c7baf7e22c49027f929729dec376"} + { + "name": "inductor_compile", + "ts": 1733527101197129.0, + "args": { + "fn_name": "compile_fx_inner", + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.206000 1500233 torch/_inductor/compile_fx.py:835] {"artifact": {"name": "fx_graph_runnable", "encoding": "string"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2470f3342eae7074d57bd8d2177ff99c"} + + import torch + from torch import tensor, device + import torch.fx as fx + from torch._dynamo.testing import rand_strided + from math import inf + import torch._inductor.inductor_prims + + import torch._dynamo.config + import torch._inductor.config + import torch._functorch.config + import torch.fx.experimental._config + torch._dynamo.config.traceable_tensor_subclasses = set() + torch._dynamo.config.suppress_errors = False + torch._dynamo.config._ddp_optimization_mode = ['ddp_optimizer', 'python_reducer', 'python_reducer_without_compiled_forward', 'no_optimization'] + torch._dynamo.config.raise_on_ctx_manager_usage = True + torch._dynamo.config._save_config_ignore = {'constant_functions', 'skipfiles_inline_module_allowlist', 'repro_after', 'repro_level'} + torch._dynamo.config.log_compilation_metrics = False + torch._dynamo.config.reorderable_logging_functions = set() + torch._inductor.config.debug = True + torch._inductor.config.fx_graph_cache = True + torch._inductor.config.pre_grad_fusion_options = {} + torch._inductor.config.post_grad_fusion_options = {} + torch._inductor.config.fx_passes_numeric_check = {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True} + torch._inductor.config.reorder_for_compute_comm_overlap_passes = ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms'] + torch._inductor.config.implicit_fallbacks = True + torch._inductor.config.debug_index_asserts = True + torch._inductor.config._fuse_ddp_communication_passes = ['fuse_ddp_with_concat_op', 'schedule_comm_wait'] + torch._inductor.config.generate_intermediate_hooks = True + torch._inductor.config.cpp.min_chunk_size = 1 + torch._inductor.config.triton.autotune_pointwise = False + torch._inductor.config.aot_inductor.metadata = {} + torch._inductor.config.aot_inductor.presets = {} + torch._inductor.config.rocm.arch = [] + torch._inductor.config.rocm.ck_supported_arch = ['gfx90a', 'gfx940', 'gfx941', 'gfx942'] + torch._inductor.config._save_config_ignore = ['trace.upload_tar', 'joint_custom_pre_pass', 'joint_custom_post_pass', 'pre_grad_custom_pass'] + torch._inductor.config._cache_config_ignore_prefix = ['trace', 'cuda.cutlass_dir', 'worker_start_method', 'compile_threads', 'post_grad_custom_post_pass', 'post_grad_custom_pre_pass', 'always_complex_memory_overlap_TESTING_ONLY'] + torch._inductor.config.external_matmul = [] + torch._functorch.config.functionalize_rng_ops = False + torch._functorch.config.enable_autograd_cache = True + torch._functorch.config.fake_tensor_allow_unsafe_data_ptr_access = True + torch._functorch.config.unlift_effect_tokens = True + + + + isolate_fails_code_str = None + + + + + # torch version: 2.6.0a0+giteece9ec + # torch cuda version: 12.2 + # torch git version: eece9ecd62cae84bc2f915fc48cffe43e30256aa + + + # CUDA Info: + # nvcc: NVIDIA (R) Cuda compiler driver + # Copyright (c) 2005-2023 NVIDIA Corporation + # Built on Tue_Aug_15_22:02:13_PDT_2023 + # Cuda compilation tools, release 12.2, V12.2.140 + # Build cuda_12.2.r12.2/compiler.33191640_0 + + # GPU Hardware Info: + # NVIDIA H100 : 8 + + + from torch.nn import * + class Repro(torch.nn.Module): + def __init__(self) -> None: + super().__init__() + + + + def forward(self, arg0_1): + mul = torch.ops.aten.mul.Tensor(arg0_1, 0.5) + mul_1 = torch.ops.aten.mul.Tensor(arg0_1, 0.7071067811865476); arg0_1 = None + erf = torch.ops.aten.erf.default(mul_1); mul_1 = None + add = torch.ops.aten.add.Tensor(erf, 1); erf = None + mul_2 = torch.ops.aten.mul.Tensor(mul, add); mul = add = None + inductor_seeds_default = torch.ops.prims.inductor_seeds.default(1, device(type='cuda', index=0)) + inductor_lookup_seed_default = torch.ops.prims.inductor_lookup_seed.default(inductor_seeds_default, 0); inductor_seeds_default = None + inductor_random_default = torch.ops.prims.inductor_random.default([1, 320, 128, 128], inductor_lookup_seed_default, 'rand'); inductor_lookup_seed_default = None + clone = torch.ops.aten.clone.default(inductor_random_default, memory_format = torch.channels_last); inductor_random_default = None + gt = torch.ops.aten.gt.Scalar(clone, 0.1); clone = None + mul_3 = torch.ops.aten.mul.Tensor(gt, mul_2); gt = mul_2 = None + mul_4 = torch.ops.aten.mul.Tensor(mul_3, 1.1111111111111112); mul_3 = None + baz = torch.ops.test.baz.default(mul_4); mul_4 = None + return (baz,) + def load_args(reader): + buf0 = reader.storage(None, 20971520, device=device(type='cuda', index=0)) + reader.tensor(buf0, (1, 320, 128, 128), (5242880, 1, 40960, 320), is_leaf=True) # arg0_1 + load_args._version = 0 + mod = Repro() + if __name__ == '__main__': + from torch._dynamo.repro.after_aot import run_repro + with torch.no_grad(): + run_repro(mod, load_args, accuracy=False, command='run', save_dir=None, tracing_mode='real', check_str=None) + # To run it separately, do + # mod, args = run_repro(mod, load_args, accuracy=False, command='get_args', save_dir=None, tracing_mode='real', check_str=None) + # mod(*args) +V1206 15:18:21.212000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "5c03608cdc5fd6af674af56b678a82ad"} + { + "name": "_recursive_post_grad_passes", + "ts": 1733527101212025.5, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.229000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "02fcd8096aaa1296b71ae553fbca81c1"} + { + "name": "_recursive_post_grad_passes", + "ts": 1733527101229240.8, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.230000 1500233 torch/_inductor/compile_fx.py:898] {"inductor_post_grad_graph": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "b1c8a699812350437e2781e8ae0cdd78"} + class (torch.nn.Module): + def forward(self, arg0_1: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0"): # No stacktrace found for following nodes - inductor_seeds_default: "i64[1]" = torch.ops.prims.inductor_seeds.default(1, device(type='cuda', index=0)) - inductor_lookup_seed_default: "i64[]" = torch.ops.prims.inductor_lookup_seed.default(inductor_seeds_default, 0); inductor_seeds_default = None - inductor_random_default: "f32[1, 320, 128, 128]" = torch.ops.prims.inductor_random.default([1, 320, 128, 128], inductor_lookup_seed_default, 'rand'); inductor_lookup_seed_default = None + inductor_seeds_default: "i64[1][1]cuda:0" = torch.ops.prims.inductor_seeds.default(1, device(type='cuda', index=0)) + inductor_lookup_seed_default: "i64[][]cuda:0" = torch.ops.prims.inductor_lookup_seed.default(inductor_seeds_default, 0); inductor_seeds_default = None + inductor_random_default: "f32[1, 320, 128, 128][5242880, 16384, 128, 1]cuda:0" = torch.ops.prims.inductor_random.default([1, 320, 128, 128], inductor_lookup_seed_default, 'rand'); inductor_lookup_seed_default = None + + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11120 in helper, code: out = self.in_layers(out) + clone: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.clone.default(inductor_random_default, memory_format = torch.channels_last); inductor_random_default = None + gt: "b8[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.gt.Scalar(clone, 0.1); clone = None + + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11119 in helper, code: out = F.gelu(x) + mul: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(arg0_1, 0.5) + mul_1: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(arg0_1, 0.7071067811865476); arg0_1 = None + erf: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.erf.default(mul_1); mul_1 = None + add: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.add.Tensor(erf, 1); erf = None + mul_2: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(mul, add); mul = add = None - # File: /data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py:8810 in helper, code: out = self.in_layers(out) - clone: "f32[1, 320, 128, 128]" = torch.ops.aten.clone.default(inductor_random_default, memory_format = torch.channels_last); inductor_random_default = None - gt: "b8[1, 320, 128, 128]" = torch.ops.aten.gt.Scalar(clone, 0.1); clone = None - mul_3: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(gt, mul_2); gt = mul_2 = None - mul_4: "f32[1, 320, 128, 128]" = torch.ops.aten.mul.Tensor(mul_3, 1.1111111111111112); mul_3 = None + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11120 in helper, code: out = self.in_layers(out) + mul_3: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(gt, mul_2); gt = mul_2 = None + mul_4: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.aten.mul.Tensor(mul_3, 1.1111111111111112); mul_3 = None - # File: /data/users/jjwu/a/pytorch/test/inductor/test_torchinductor.py:8815 in forward, code: out = torch.ops.test.baz(out) - baz: "f32[1, 320, 128, 128]" = torch.ops.test.baz.default(mul_4); mul_4 = None + # File: /data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py:11125 in forward, code: out = torch.ops.test.baz(out) + baz: "f32[1, 320, 128, 128][5242880, 1, 40960, 320]cuda:0" = torch.ops.test.baz.default(mul_4); mul_4 = None return (baz,) -V0401 08:54:22.975000 140424060892160 torch/_inductor/graph.py:1268] {"inductor_output_code": {"filename": "/tmp/torchinductor_jjwu/pz/cpzf3cxhhnoarj4kjhg2wxhmski6yge4zox6h43vsrrdubhj7qnb.py"}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "5dd8b94f459b07b5314187f78575d118"} - - from ctypes import c_void_p, c_long +V1206 15:18:21.247000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "cd84113519eaf110be2573a6d9ab7d5c"} + { + "name": "GraphLowering.run", + "ts": 1733527101247266.8, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.284000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "8c60f183722282266a91df7c4e6a649f"} + { + "name": "GraphLowering.run", + "ts": 1733527101284562.5, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.284000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "170214c32ec5bf03937f621741046dcd"} + { + "name": "GraphLowering.compile_to_fn", + "ts": 1733527101284939.5, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.285000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2224b3ed7a31123f4b48c477e08d0cd1"} + { + "name": "code_gen", + "ts": 1733527101285145.2, + "args": { + "fn_name": "GraphLowering.compile_to_module", + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.285000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "1f11a0e6ad50d55bb5656736d0b49cb8"} + { + "name": "GraphLowering.codegen", + "ts": 1733527101285316.5, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.287000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "6644f6874ee64c9e36e4f40bfd6f3d1a"} + { + "name": "Scheduler.__init__", + "ts": 1733527101287676.0, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.304000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "bd092078f3a275d331fa39b88de2d12c"} + { + "name": "Scheduler.fused_nodes", + "ts": 1733527101304485.8, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.305000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "41a7f1db6eb37c96544c9ab2fda60033"} + { + "name": "Scheduler.fused_nodes", + "ts": 1733527101305930.5, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.308000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "4aa4bc04054bdb300303a01c86fdbf07"} + { + "name": "Scheduler.__init__", + "ts": 1733527101308460.5, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.308000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "3548e38bf922ab2f500ca438d9dd5db4"} + { + "name": "Scheduler.codegen", + "ts": 1733527101308690.8, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.450000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "cb811b477878494e7cee5bcffb0dd0fa"} + { + "name": "Scheduler.codegen", + "ts": 1733527101450232.0, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.450000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "d3cf9b88610841c93e68e258b4160410"} + { + "name": "PythonWrapperCodegen.generate", + "ts": 1733527101450612.8, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.451000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "12933d2f0c133902de1bcf9876e27aa4"} + { + "name": "PythonWrapperCodegen.generate", + "ts": 1733527101451570.0, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.451000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "8661f4a66b85846d47bc578e43957bb5"} + { + "name": "GraphLowering.codegen", + "ts": 1733527101451789.8, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.452000 1500233 torch/_inductor/graph.py:2030] {"inductor_output_code": {"filename": "/tmp/tmprds_hch0/yj/cyjwx6x6efpuwt4dvr4ev42v4ghac5zo2uggvscht2otwnesito6.py"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "35842d763c8905794995d6c4a089d3b3"} + # AOT ID: ['0_inference'] + from ctypes import c_void_p, c_long, c_int import torch import math import random @@ -84,49 +638,65 @@ V0401 08:54:22.975000 140424060892160 torch/_inductor/graph.py:1268] {"inductor_ from torch._inductor.hooks import run_intermediate_hooks from torch._inductor.utils import maybe_profile from torch._inductor.codegen.memory_planning import _align as align - from torch import device, empty_strided - from torch._inductor.codecache import AsyncCompile + from torch._inductor.async_compile import AsyncCompile from torch._inductor.select_algorithm import extern_kernels from torch._inductor.codegen.multi_kernel import MultiKernelCall + import triton + import triton.language as tl + from torch._inductor.runtime.triton_heuristics import ( + grid, + split_scan_grid, + grid_combo_kernels, + start_graph, + end_graph, + cooperative_reduction_grid, + ) + from torch._C import _cuda_getCurrentRawStream as get_raw_stream aten = torch.ops.aten inductor_ops = torch.ops.inductor + _quantized = torch.ops._quantized assert_size_stride = torch._C._dynamo.guards.assert_size_stride empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda + empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu + reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor alloc_from_pool = torch.ops.inductor._alloc_from_pool - reinterpret_tensor = torch.ops.inductor._reinterpret_tensor async_compile = AsyncCompile() + empty_strided_p2p = torch._C._distributed_c10d._SymmetricMemory.empty_strided_p2p - # kernel path: /tmp/torchinductor_jjwu/6p/c6pjjivx4yr7a6rhwljusernnkewuf53hvlnc2sy63rcnj764rhn.py - # Source Nodes: [], Original ATen: [] - - triton_poi_fused_0 = async_compile.triton('triton_', ''' + # kernel path: /tmp/tmprds_hch0/ke/ckedh2vjam5uo7wobyr5yq2et3clblzbzgykujgmjbmkj5uyimpl.py + # Topologically Sorted Source Nodes: [input_1], Original ATen: [aten.native_dropout] + # Source node to ATen node mapping: + # input_1 => inductor_lookup_seed_default, inductor_random_default + # Graph fragment: + # %inductor_lookup_seed_default : [num_users=1] = call_function[target=torch.ops.prims.inductor_lookup_seed.default](args = (%inductor_seeds_default, 0), kwargs = {}) + # %inductor_random_default : [num_users=1] = call_function[target=torch.ops.prims.inductor_random.default](args = ([1, 320, 128, 128], %inductor_lookup_seed_default, rand), kwargs = {}) + triton_poi_fused_native_dropout_0 = async_compile.triton('triton_poi_fused_native_dropout_0', ''' import triton import triton.language as tl from triton.compiler.compiler import AttrsDescriptor - from torch._inductor import triton_helpers, triton_heuristics - from torch._inductor.ir import ReductionHint, TileHint - from torch._inductor.triton_helpers import libdevice, math as tl_math - from torch._inductor.triton_heuristics import AutotuneHint - from torch._inductor.utils import instance_descriptor + from torch._inductor.runtime import triton_helpers, triton_heuristics + from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math + from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, DeviceProperties + triton_helpers.set_driver_to_gpu() @triton_heuristics.pointwise( size_hints=[8388608], filename=__file__, - triton_meta={'signature': {0: '*i64', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': 0, 'device_type': 'cuda', 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 3), equal_to_1=(), divisible_by_8=(3,))]}, - inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_0', 'mutated_arg_names': [], 'no_x_dim': False, 'backend_hash': 'e24e28e8c74b85ff7b61b41fc9160c05d25c32556bda76a915743727cec50966'}, + triton_meta={'signature': {'in_ptr0': '*i64', 'out_ptr0': '*fp32', 'load_seed_offset': 'i32', 'xnumel': 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=90, major=9, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=132, warp_size=32), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 3), equal_to_1=())]}, + inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_dropout_0', 'mutated_arg_names': [], 'optimize_mem': True, 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': '562E840D41CEB1D8E51DE726EA7592B0C37A0C6FBD72CF4E958863CEC11D41A7', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': False, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}, min_elem_per_thread=0 ) @triton.jit - def triton_(in_ptr0, out_ptr0, load_seed_offset, xnumel, XBLOCK : tl.constexpr): + def triton_poi_fused_native_dropout_0(in_ptr0, out_ptr0, load_seed_offset, xnumel, XBLOCK : tl.constexpr): xnumel = 5242880 xoffset = tl.program_id(0) * XBLOCK xindex = xoffset + tl.arange(0, XBLOCK)[:] - xmask = xindex < xnumel + xmask = tl.full([XBLOCK], True, tl.int1) x0 = xindex tmp0 = tl.load(in_ptr0 + load_seed_offset) tmp1 = x0 @@ -134,49 +704,55 @@ V0401 08:54:22.975000 140424060892160 torch/_inductor/graph.py:1268] {"inductor_ tl.store(out_ptr0 + (x0), tmp2, None) ''', device_str='cuda') - import triton - import triton.language as tl - from torch._inductor.triton_heuristics import grid, split_scan_grid, start_graph, end_graph - from torch._C import _cuda_getCurrentRawStream as get_raw_stream - - # kernel path: /tmp/torchinductor_jjwu/2g/c2gsdapii4jxiorppiwpmvslqswzpbmms26vhuqoh7nm27gxhdnf.py - # Source Nodes: [out, out_2, out_3], Original ATen: [aten.gelu, aten.native_dropout, test.baz] - # out => add, erf, mul, mul_1, mul_2 - # out_2 => clone, gt, mul_3, mul_4 - # out_3 => baz - triton_poi_fused_baz_gelu_native_dropout_1 = async_compile.triton('triton_', ''' + # kernel path: /tmp/tmprds_hch0/eu/ceutejflq32k5wvvsucbkscrxmvjorlj2t3eq7tgcfyqrz5mhnh3.py + # Topologically Sorted Source Nodes: [input_1, out, out_1], Original ATen: [aten.native_dropout, aten.gelu, test.baz] + # Source node to ATen node mapping: + # input_1 => clone, gt, mul_3, mul_4 + # out => add, erf, mul, mul_1, mul_2 + # out_1 => baz + # Graph fragment: + # %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%inductor_random_default,), kwargs = {memory_format: torch.channels_last}) + # %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%clone, 0.1), kwargs = {}) + # %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, 0.5), kwargs = {}) + # %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, 0.7071067811865476), kwargs = {}) + # %erf : [num_users=1] = call_function[target=torch.ops.aten.erf.default](args = (%mul_1,), kwargs = {}) + # %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%erf, 1), kwargs = {}) + # %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %add), kwargs = {}) + # %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%gt, %mul_2), kwargs = {}) + # %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_3, 1.1111111111111112), kwargs = {}) + # %baz : [num_users=1] = call_function[target=torch.ops.test.baz.default](args = (%mul_4,), kwargs = {}) + triton_poi_fused_baz_gelu_native_dropout_1 = async_compile.triton('triton_poi_fused_baz_gelu_native_dropout_1', ''' import triton import triton.language as tl from triton.compiler.compiler import AttrsDescriptor - from torch._inductor import triton_helpers, triton_heuristics - from torch._inductor.ir import ReductionHint, TileHint - from torch._inductor.triton_helpers import libdevice, math as tl_math - from torch._inductor.triton_heuristics import AutotuneHint - from torch._inductor.utils import instance_descriptor + from torch._inductor.runtime import triton_helpers, triton_heuristics + from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math + from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, DeviceProperties + triton_helpers.set_driver_to_gpu() @triton_heuristics.pointwise( size_hints=[16384, 512], tile_hint=TileHint.DEFAULT, filename=__file__, - triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': 0, 'device_type': 'cuda', 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=(), divisible_by_8=(3, 4))]}, - inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_baz_gelu_native_dropout_1', 'mutated_arg_names': [], 'no_x_dim': False, 'backend_hash': 'e24e28e8c74b85ff7b61b41fc9160c05d25c32556bda76a915743727cec50966'}, + triton_meta={'signature': {'in_ptr0': '*fp32', 'in_ptr1': '*fp32', 'out_ptr0': '*fp32', 'ynumel': 'i32', 'xnumel': 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=90, major=9, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=132, warp_size=32), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]}, + inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_baz_gelu_native_dropout_1', 'mutated_arg_names': [], 'optimize_mem': True, 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': '562E840D41CEB1D8E51DE726EA7592B0C37A0C6FBD72CF4E958863CEC11D41A7', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': False, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}, min_elem_per_thread=0 ) @triton.jit - def triton_(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr): + def triton_poi_fused_baz_gelu_native_dropout_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr): ynumel = 16384 xnumel = 320 - yoffset = tl.program_id(1) * (tl.program_id(2) + 1) * YBLOCK + yoffset = tl.program_id(1) * YBLOCK yindex = yoffset + tl.arange(0, YBLOCK)[None, :] - ymask = yindex < ynumel + ymask = tl.full([XBLOCK, YBLOCK], True, tl.int1) xoffset = tl.program_id(0) * XBLOCK xindex = xoffset + tl.arange(0, XBLOCK)[:, None] xmask = xindex < xnumel x1 = xindex y0 = yindex - tmp0 = tl.load(in_ptr0 + (y0 + (16384*x1)), xmask, eviction_policy='evict_last') - tmp4 = tl.load(in_ptr1 + (x1 + (320*y0)), xmask, eviction_policy='evict_last') + tmp0 = tl.load(in_ptr0 + (y0 + 16384*x1), xmask, eviction_policy='evict_last') + tmp4 = tl.load(in_ptr1 + (x1 + 320*y0), xmask, eviction_policy='evict_last') tmp1 = 0.1 tmp2 = tmp0 > tmp1 tmp3 = tmp2.to(tl.float32) @@ -191,7 +767,7 @@ V0401 08:54:22.975000 140424060892160 torch/_inductor/graph.py:1268] {"inductor_ tmp13 = tmp3 * tmp12 tmp14 = 1.1111111111111112 tmp15 = tmp13 * tmp14 - tl.store(out_ptr0 + (x1 + (320*y0)), tmp15, xmask) + tl.store(out_ptr0 + (x1 + 320*y0), tmp15, xmask) ''', device_str='cuda') @@ -205,21 +781,19 @@ V0401 08:54:22.975000 140424060892160 torch/_inductor/graph.py:1268] {"inductor_ with torch.cuda._DeviceGuard(0): torch.cuda.set_device(0) buf0 = empty_strided_cuda((1, ), (1, ), torch.int64) - # Source Nodes: [], Original ATen: [] + # Topologically Sorted Source Nodes: [], Original ATen: [] aten.randint.low_out(-9223372036854775808, 9223372036854775807, [1], out=buf0) buf1 = empty_strided_cuda((1, 320, 128, 128), (5242880, 16384, 128, 1), torch.float32) - # Source Nodes: [], Original ATen: [] + # Topologically Sorted Source Nodes: [input_1], Original ATen: [aten.native_dropout] stream0 = get_raw_stream(0) - triton_poi_fused_0.run(buf0, buf1, 0, 5242880, grid=grid(5242880), stream=stream0) - run_intermediate_hooks('inductor_random_default', buf1) + triton_poi_fused_native_dropout_0.run(buf0, buf1, 0, 5242880, grid=grid(5242880), stream=stream0) del buf0 buf2 = empty_strided_cuda((1, 320, 128, 128), (5242880, 1, 40960, 320), torch.float32) - # Source Nodes: [out, out_2, out_3], Original ATen: [aten.gelu, aten.native_dropout, test.baz] + # Topologically Sorted Source Nodes: [input_1, out, out_1], Original ATen: [aten.native_dropout, aten.gelu, test.baz] triton_poi_fused_baz_gelu_native_dropout_1.run(buf1, arg0_1, buf2, 16384, 320, grid=grid(16384, 320), stream=stream0) - run_intermediate_hooks('mul_4', buf2) del arg0_1 del buf1 - # Source Nodes: [out, out_2, out_3], Original ATen: [aten.gelu, aten.native_dropout, test.baz] + # Topologically Sorted Source Nodes: [input_1, out, out_1], Original ATen: [aten.native_dropout, aten.gelu, test.baz] buf3 = torch.ops.test.baz.default(buf2) run_intermediate_hooks('baz', buf3) del buf2 @@ -240,19 +814,1411 @@ V0401 08:54:22.975000 140424060892160 torch/_inductor/graph.py:1268] {"inductor_ from torch._inductor.wrapper_benchmark import compiled_module_main compiled_module_main('None', benchmark_compiled_module) -V0401 08:54:22.980000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/convert_frame.py", 11]} -V0401 08:54:22.980000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/utils.py", 12]} -V0401 08:54:22.980000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/bytecode_transformation.py", 13]} -V0401 08:54:22.980000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/symbolic_convert.py", 14]} -V0401 08:54:22.980000 140424060892160 torch/_logging/structured.py:19] {"str": ["/data/users/jjwu/a/pytorch/torch/_dynamo/output_graph.py", 15]} -V0401 08:54:22.980000 140424060892160 torch/_dynamo/guards.py:1194] {"dynamo_guards": {}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2fd53e7b434e73e7bbdc496799a81b21"} - [ - {"code": "hasattr(L['x'], '_dynamo_dynamic_indices') == False", "stack": null, "user_stack": null}, - {"code": "___check_obj_id(L['self'], 140423721840032)", "stack": null, "user_stack": null}, - {"code": "___check_obj_id(L['self'].training, 7665376)", "stack": null, "user_stack": null}, - {"code": "utils_device.CURRENT_DEVICE == None", "stack": [{"line": 10031, "name": "", "filename": 0}, {"line": 14, "name": "run_tests", "filename": 1}, {"line": 41, "name": "run_tests", "filename": 2}, {"line": 1165, "name": "run_tests", "filename": 3}, {"line": 101, "name": "__init__", "filename": 4}, {"line": 271, "name": "runTests", "filename": 4}, {"line": 184, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 650, "name": "__call__", "filename": 7}, {"line": 2866, "name": "run", "filename": 3}, {"line": 2838, "name": "_run_custom", "filename": 3}, {"line": 591, "name": "run", "filename": 7}, {"line": 549, "name": "_callTestMethod", "filename": 7}, {"line": 2739, "name": "wrapper", "filename": 3}, {"line": 9214, "name": "new_test", "filename": 0}, {"line": 79, "name": "inner", "filename": 8}, {"line": 8845, "name": "test_custom_op_fixed_layout_channels_last", "filename": 0}, {"line": 1527, "name": "_wrapped_call_impl", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}, {"line": 450, "name": "_fn", "filename": 10}, {"line": 1527, "name": "_wrapped_call_impl", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}, {"line": 939, "name": "catch_errors", "filename": 11}, {"line": 802, "name": "_convert_frame", "filename": 11}, {"line": 400, "name": "_convert_frame_assert", "filename": 11}, {"line": 79, "name": "inner", "filename": 8}, {"line": 686, "name": "_compile", "filename": 11}, {"line": 262, "name": "time_wrapper", "filename": 12}, {"line": 541, "name": "compile_inner", "filename": 11}, {"line": 1036, "name": "transform_code_object", "filename": 13}, {"line": 165, "name": "_fn", "filename": 11}, {"line": 485, "name": "transform", "filename": 11}, {"line": 2105, "name": "__init__", "filename": 14}, {"line": 344, "name": "__init__", "filename": 15}, {"line": 467, "name": "init_ambient_guards", "filename": 15}], "user_stack": null}, - {"code": "___check_current_backend(140423721840560)", "stack": [{"line": 10031, "name": "", "filename": 0}, {"line": 14, "name": "run_tests", "filename": 1}, {"line": 41, "name": "run_tests", "filename": 2}, {"line": 1165, "name": "run_tests", "filename": 3}, {"line": 101, "name": "__init__", "filename": 4}, {"line": 271, "name": "runTests", "filename": 4}, {"line": 184, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 650, "name": "__call__", "filename": 7}, {"line": 2866, "name": "run", "filename": 3}, {"line": 2838, "name": "_run_custom", "filename": 3}, {"line": 591, "name": "run", "filename": 7}, {"line": 549, "name": "_callTestMethod", "filename": 7}, {"line": 2739, "name": "wrapper", "filename": 3}, {"line": 9214, "name": "new_test", "filename": 0}, {"line": 79, "name": "inner", "filename": 8}, {"line": 8845, "name": "test_custom_op_fixed_layout_channels_last", "filename": 0}, {"line": 1527, "name": "_wrapped_call_impl", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}, {"line": 450, "name": "_fn", "filename": 10}, {"line": 1527, "name": "_wrapped_call_impl", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}, {"line": 939, "name": "catch_errors", "filename": 11}, {"line": 802, "name": "_convert_frame", "filename": 11}, {"line": 400, "name": "_convert_frame_assert", "filename": 11}, {"line": 79, "name": "inner", "filename": 8}, {"line": 686, "name": "_compile", "filename": 11}, {"line": 262, "name": "time_wrapper", "filename": 12}, {"line": 541, "name": "compile_inner", "filename": 11}, {"line": 1036, "name": "transform_code_object", "filename": 13}, {"line": 165, "name": "_fn", "filename": 11}, {"line": 485, "name": "transform", "filename": 11}, {"line": 2105, "name": "__init__", "filename": 14}, {"line": 344, "name": "__init__", "filename": 15}, {"line": 473, "name": "init_ambient_guards", "filename": 15}], "user_stack": null}, - {"code": "check_tensor(L['x'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 320, 128, 128], stride=[5242880, 1, 40960, 320])", "stack": null, "user_stack": null} - ] -V0401 08:54:21.882000 140424060892160 torch/_dynamo/convert_frame.py:672] {"dynamo_start": {"stack": [{"line": 10031, "name": "", "filename": 0}, {"line": 14, "name": "run_tests", "filename": 1}, {"line": 41, "name": "run_tests", "filename": 2}, {"line": 1165, "name": "run_tests", "filename": 3}, {"line": 101, "name": "__init__", "filename": 4}, {"line": 271, "name": "runTests", "filename": 4}, {"line": 184, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 650, "name": "__call__", "filename": 7}, {"line": 2866, "name": "run2", "filename": 3}, {"line": 2838, "name": "_run_custom", "filename": 3}, {"line": 591, "name": "run", "filename": 7}, {"line": 549, "name": "_callTestMethod", "filename": 7}, {"line": 2739, "name": "wrapper", "filename": 3}, {"line": 9214, "name": "new_test", "filename": 0}, {"line": 79, "name": "inner", "filename": 8}, {"line": 8845, "name": "test_custom_op_fixed_layout_channels_last", "filename": 0}, {"line": 1527, "name": "_wrapped_call_impl", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}, {"line": 450, "name": "_fn", "filename": 10}, {"line": 1527, "name": "_wrapped_call_impl", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}]}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} -V0401 08:54:21.882000 140424060892160 torch/_dynamo/convert_frame.py:672] {"dynamo_start": {"stack": [{"line": 10031, "name": "", "filename": 0}, {"line": 14, "name": "run_tests", "filename": 1}, {"line": 41, "name": "run_tests", "filename": 2}, {"line": 1165, "name": "run_tests", "filename": 3}, {"line": 101, "name": "__init__", "filename": 4}, {"line": 271, "name": "runTests", "filename": 4}, {"line": 184, "name": "run", "filename": 5}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 84, "name": "__call__", "filename": 6}, {"line": 122, "name": "run", "filename": 6}, {"line": 650, "name": "__call__", "filename": 7}, {"line": 2866, "name": "run2", "filename": 3}, {"line": 2838, "name": "_run_custom", "filename": 3}, {"line": 591, "name": "run", "filename": 7}, {"line": 549, "name": "_callTestMethod", "filename": 7}, {"line": 2739, "name": "wrapper", "filename": 3}, {"line": 9214, "name": "new_test", "filename": 0}, {"line": 79, "name": "inner", "filename": 8}, {"line": 8845, "name": "test_custom_op_fixed_layout_channels_last", "filename": 0}, {"line": 1527, "name": "_wrapped_call_impl", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}, {"line": 450, "name": "_fn", "filename": 10}, {"line": 1527, "name": "_wrapped_call_implaa", "filename": 9}, {"line": 1536, "name": "_call_impl", "filename": 9}]}, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:18:21.452000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "e0fd41277ac80fe5bd868c65d517d89b"} + { + "name": "PyCodeCache.load_by_key_path", + "ts": 1733527101452695.0, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.461000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "006d0abc072724b3a8ab6d84182d8431"} + { + "name": "async_compile.wait", + "ts": 1733527101461170.2, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.813000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "50c5a1c818507a194263d8d974159dd7"} + { + "name": "async_compile.wait", + "ts": 1733527101813608.5, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.814000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2afa4a7e52f0e3b41010122154a8348f"} + { + "name": "PyCodeCache.load_by_key_path", + "ts": 1733527101814085.2, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.814000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "ee8f2db95d5b4f7e8cca64df4854e627"} + { + "name": "code_gen", + "ts": 1733527101814319.8, + "args": { + "fn_name": "GraphLowering.compile_to_module", + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.814000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "600f56ac93eb73f3fb979f2e5d61dc44"} + { + "name": "GraphLowering.compile_to_fn", + "ts": 1733527101814515.0, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.900000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "ff8853e249b1cd98d1a1c852aeb5bddd"} + { + "name": "TritonBundler.collect", + "ts": 1733527101900813.5, + "args": { + "compile_id": "0/0" + }, + "ph": "B", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.901000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "16e2d9e68e2009468c38b0d573810760"} + { + "name": "TritonBundler.collect", + "ts": 1733527101901684.5, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.902000 1500233 torch/_dynamo/utils.py:1327] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "44378945e331d48456c66d308783acfa"} + { + "name": "fx_graph_cache_miss", + "ts": 1733527101197944.0, + "args": { + "key": "ftyqixqyigk4dzwkgsofqaadk6duwj57z5bipzhn5alvsznuf2m7", + "components": [ + "[pvmcgcpspyiotfrhpzur22xkrqhko5vx57txiq4wghhk7pm3ocb] gm: ()\n\n\n\ndef forward(self, arg0_1):\n mul = torch.ops.aten.mul.Tensor(arg0_1, 0.5)\n mul_1 = torch.ops.aten.mul.Tensor(arg0_1, 0.7071067811865476); arg0_1 = None\n erf = torch.ops.aten.erf.default(mul_1); mul_1 = None\n add = torch.ops.aten.add.Tensor(erf, 1); erf = None\n mul_2 = torch.ops.aten.mul.Tensor(mul, add); mul = add = None\n inductor_seeds_default = torch.ops.prims.inductor_seeds.default(1, device(type='cuda', index=0))\n inductor_lookup_seed_default = torch.ops.prims.inductor_lookup_seed.default(inductor_seeds_default, 0); inductor_seeds_default = None\n inductor_random_default = torch.ops.prims.inductor_random.default([1, 320, 128, 128], inductor_lookup_seed_default, 'rand'); inductor_lookup_seed_default = None\n clone = torch.ops.aten.clone.default(inductor_random_default, memory_format = torch.channels_last); inductor_random_default = None\n gt = torch.ops.aten.gt.Scalar(clone, 0.1); clone = None\n mul_3 = torch.ops.aten.mul.Tensor(gt, mul_2); gt = mul_2 = None\n mul_4 = torch.ops.aten.mul.Tensor(mul_3, 1.1111111111111112); mul_3 = None\n baz = torch.ops.test.baz.default(mul_4); mul_4 = None\n return (baz,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[qqiqyf6plgvsceqtyjt7qfdyz4oo3oixqvjcpsmypqbyyipyaif] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 320, 128, 128]), stride=(5242880, 1, 40960, 320), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.channels_last, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", + "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", + "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", + "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", + "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[fx_graph_remote_cache]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_local_cache]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[autotune_remote_cache]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[bundled_autotune_remote_cache]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", + "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", + "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", + "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", + "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug_index_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", + "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[generate_intermediate_hooks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.min_chunk_size]: 1", + "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.autotune_pointwise]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", + "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" + ], + "cache_event_time": 1733527101203264662, + "cache_state": "miss", + "triton_bundler_meta": "TritonBundlerMetadata(cached_kernel_names=['triton_poi_fused_native_dropout_0', 'triton_poi_fused_baz_gelu_native_dropout_1'])", + "time_taken_ns": 702837316, + "compile_id": "0/0" + }, + "ph": "i", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0, + "s": "p" + } +V1206 15:18:21.903000 1500233 torch/_inductor/compile_fx.py:751] {"artifact": {"name": "fx_graph_cache_miss", "encoding": "json"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "bd34f8d5d775fcaf77d2c02d265bee67"} + {"key": "ftyqixqyigk4dzwkgsofqaadk6duwj57z5bipzhn5alvsznuf2m7", "components": ["[pvmcgcpspyiotfrhpzur22xkrqhko5vx57txiq4wghhk7pm3ocb] gm: ()\n\n\n\ndef forward(self, arg0_1):\n mul = torch.ops.aten.mul.Tensor(arg0_1, 0.5)\n mul_1 = torch.ops.aten.mul.Tensor(arg0_1, 0.7071067811865476); arg0_1 = None\n erf = torch.ops.aten.erf.default(mul_1); mul_1 = None\n add = torch.ops.aten.add.Tensor(erf, 1); erf = None\n mul_2 = torch.ops.aten.mul.Tensor(mul, add); mul = add = None\n inductor_seeds_default = torch.ops.prims.inductor_seeds.default(1, device(type='cuda', index=0))\n inductor_lookup_seed_default = torch.ops.prims.inductor_lookup_seed.default(inductor_seeds_default, 0); inductor_seeds_default = None\n inductor_random_default = torch.ops.prims.inductor_random.default([1, 320, 128, 128], inductor_lookup_seed_default, 'rand'); inductor_lookup_seed_default = None\n clone = torch.ops.aten.clone.default(inductor_random_default, memory_format = torch.channels_last); inductor_random_default = None\n gt = torch.ops.aten.gt.Scalar(clone, 0.1); clone = None\n mul_3 = torch.ops.aten.mul.Tensor(gt, mul_2); gt = mul_2 = None\n mul_4 = torch.ops.aten.mul.Tensor(mul_3, 1.1111111111111112); mul_3 = None\n baz = torch.ops.test.baz.default(mul_4); mul_4 = None\n return (baz,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[qqiqyf6plgvsceqtyjt7qfdyz4oo3oixqvjcpsmypqbyyipyaif] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 320, 128, 128]), stride=(5242880, 1, 40960, 320), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.channels_last, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[fx_graph_remote_cache]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_local_cache]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[autotune_remote_cache]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[bundled_autotune_remote_cache]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug_index_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[generate_intermediate_hooks]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.min_chunk_size]: 1", "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.autotune_pointwise]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None"], "cache_event_time": 1733527101203264662, "cache_state": "miss", "triton_bundler_meta": "TritonBundlerMetadata(cached_kernel_names=['triton_poi_fused_native_dropout_0', 'triton_poi_fused_baz_gelu_native_dropout_1'])", "time_taken_ns": 702837316, "compile_id": "0/0"} +V1206 15:18:21.903000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "785c297e1e81aef3095cf23f30840768"} + { + "name": "inductor_compile", + "ts": 1733527101903705.0, + "args": { + "fn_name": "compile_fx_inner", + "compile_id": "0/0", + "is_backward": false, + "cache_state": "miss", + "cache_event_time": 1733527101197943979, + "key": "ftyqixqyigk4dzwkgsofqaadk6duwj57z5bipzhn5alvsznuf2m7", + "components": [ + "[pvmcgcpspyiotfrhpzur22xkrqhko5vx57txiq4wghhk7pm3ocb] gm: ()\n\n\n\ndef forward(self, arg0_1):\n mul = torch.ops.aten.mul.Tensor(arg0_1, 0.5)\n mul_1 = torch.ops.aten.mul.Tensor(arg0_1, 0.7071067811865476); arg0_1 = None\n erf = torch.ops.aten.erf.default(mul_1); mul_1 = None\n add = torch.ops.aten.add.Tensor(erf, 1); erf = None\n mul_2 = torch.ops.aten.mul.Tensor(mul, add); mul = add = None\n inductor_seeds_default = torch.ops.prims.inductor_seeds.default(1, device(type='cuda', index=0))\n inductor_lookup_seed_default = torch.ops.prims.inductor_lookup_seed.default(inductor_seeds_default, 0); inductor_seeds_default = None\n inductor_random_default = torch.ops.prims.inductor_random.default([1, 320, 128, 128], inductor_lookup_seed_default, 'rand'); inductor_lookup_seed_default = None\n clone = torch.ops.aten.clone.default(inductor_random_default, memory_format = torch.channels_last); inductor_random_default = None\n gt = torch.ops.aten.gt.Scalar(clone, 0.1); clone = None\n mul_3 = torch.ops.aten.mul.Tensor(gt, mul_2); gt = mul_2 = None\n mul_4 = torch.ops.aten.mul.Tensor(mul_3, 1.1111111111111112); mul_3 = None\n baz = torch.ops.test.baz.default(mul_4); mul_4 = None\n return (baz,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[qqiqyf6plgvsceqtyjt7qfdyz4oo3oixqvjcpsmypqbyyipyaif] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 320, 128, 128]), stride=(5242880, 1, 40960, 320), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.channels_last, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[aot_mode]: False", + "[lmglpn4zi7vob56n34r2j2rk7flv5xfgrcvmo7xcpirqsitygqx] fx_kwargs[boxed_forward_device_index]: BoxedDeviceIndex(value=None)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[cpp_wrapper]: False", + "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[extern_node_serializer]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] fx_kwargs[is_backward]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] fx_kwargs[is_inference]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] fx_kwargs[layout_opt]: None", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] fx_kwargs[static_input_idxs]: []", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inputs_to_check[0]: 0", + "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", + "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[fx_graph_remote_cache]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_local_cache]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[autotune_remote_cache]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[bundled_autotune_remote_cache]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", + "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", + "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", + "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", + "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug_index_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", + "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[generate_intermediate_hooks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.min_chunk_size]: 1", + "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.autotune_pointwise]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", + "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" + ], + "cache_bypass_reason": null, + "remote_cache_enabled": false, + "local_cache_enabled": true + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.904000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "ab4c2fac9b0442f97c51db98db8465ea"} + { + "name": "compile_fx..fw_compiler_base", + "ts": 1733527101904297.0, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.906000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "b95d05f6f59e0845136feab28a56710f"} + { + "name": "create_aot_dispatcher_function", + "ts": 1733527101906249.2, + "args": { + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.906000 1500233 torch/_dynamo/utils.py:1327] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "7747dc628b14f97c4c4acd0ba40994f5"} + { + "name": "autograd_cache_miss", + "ts": 1733527101086465.2, + "args": { + "key": "avcrrszv2esbpxnxz7vvqltm5eurkzwsst2nqqq3sks3p2h3bs3t", + "cache_state": "miss", + "components": [ + "[rercb27oxulpvyy73cg2bk7544vnpfzorcoyihcqcrxerzxuiwh] aot_config: (0, True, False, False, False, [LocalSource(local_name='x', is_input=True, is_root_frame_cell=False)], True, False)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] grad_enabled: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] disable_amp: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] deterministic_algorithms: False", + "[35fcxhrk3gooceher4yasifwf7bz6x35efasz4elzmcyrazyaai] autograd_config: ", + "[nilk63auvy3gsylfusxx5tw2tokdolrygfdskikekzdsd5xiuio] gm: GraphModule()\n\n\n\ndef forward(self, L_x_ : torch.Tensor):\n l_x_ = L_x_\n out = torch._C._nn.gelu(l_x_); l_x_ = None\n input_1 = torch.nn.functional.dropout(out, 0.1, True, False); out = None\n out_1 = torch.ops.test.baz(input_1); input_1 = None\n return (out_1,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[qqiqyf6plgvsceqtyjt7qfdyz4oo3oixqvjcpsmypqbyyipyaif] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 320, 128, 128]), stride=(5242880, 1, 40960, 320), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.channels_last, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", + "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", + "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[fx_graph_remote_cache]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_local_cache]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[autotune_remote_cache]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[bundled_autotune_remote_cache]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", + "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", + "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", + "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", + "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug_index_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", + "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[generate_intermediate_hooks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.min_chunk_size]: 1", + "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.autotune_pointwise]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", + "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" + ], + "compile_id": "0/0" + }, + "ph": "i", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0, + "s": "p" + } +V1206 15:18:21.906000 1500233 torch/_functorch/_aot_autograd/autograd_cache.py:763] {"artifact": {"name": "aotautograd_cache_hash", "encoding": "json"}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "af016cc50869012432c17034d2b4b2c0"} + {"key": "avcrrszv2esbpxnxz7vvqltm5eurkzwsst2nqqq3sks3p2h3bs3t", "cache_state": "miss", "components": ["[rercb27oxulpvyy73cg2bk7544vnpfzorcoyihcqcrxerzxuiwh] aot_config: (0, True, False, False, False, [LocalSource(local_name='x', is_input=True, is_root_frame_cell=False)], True, False)", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] grad_enabled: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] disable_amp: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] deterministic_algorithms: False", "[35fcxhrk3gooceher4yasifwf7bz6x35efasz4elzmcyrazyaai] autograd_config: ", "[nilk63auvy3gsylfusxx5tw2tokdolrygfdskikekzdsd5xiuio] gm: GraphModule()\n\n\n\ndef forward(self, L_x_ : torch.Tensor):\n l_x_ = L_x_\n out = torch._C._nn.gelu(l_x_); l_x_ = None\n input_1 = torch.nn.functional.dropout(out, 0.1, True, False); out = None\n out_1 = torch.ops.test.baz(input_1); input_1 = None\n return (out_1,)\n \n# To see more debug info, please use `graph_module.print_readable()`", "[qqiqyf6plgvsceqtyjt7qfdyz4oo3oixqvjcpsmypqbyyipyaif] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 320, 128, 128]), stride=(5242880, 1, 40960, 320), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.channels_last, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[fx_graph_remote_cache]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_local_cache]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[autotune_remote_cache]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[bundled_autotune_remote_cache]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug_index_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[generate_intermediate_hooks]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.min_chunk_size]: 1", "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.autotune_pointwise]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None"], "compile_id": "0/0"} +V1206 15:18:21.907000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "210650099f3176e17ed6c62583692e36"} + { + "name": "backend_compile", + "ts": 1733527101907180.8, + "args": { + "fn_name": "OutputGraph.call_user_compiler", + "compile_id": "0/0", + "requires_subclass_dispatch": false, + "dispatch_mode": "inference", + "cache_state": "miss", + "cache_event_time": 1733527101086465244, + "key": "avcrrszv2esbpxnxz7vvqltm5eurkzwsst2nqqq3sks3p2h3bs3t", + "components": [ + "[rercb27oxulpvyy73cg2bk7544vnpfzorcoyihcqcrxerzxuiwh] aot_config: (0, True, False, False, False, [LocalSource(local_name='x', is_input=True, is_root_frame_cell=False)], True, False)", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] grad_enabled: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] disable_amp: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] deterministic_algorithms: False", + "[35fcxhrk3gooceher4yasifwf7bz6x35efasz4elzmcyrazyaai] autograd_config: ", + "[nilk63auvy3gsylfusxx5tw2tokdolrygfdskikekzdsd5xiuio] gm: GraphModule()\n\n\n\ndef forward(self, L_x_ : torch.Tensor):\n l_x_ = L_x_\n out = torch._C._nn.gelu(l_x_); l_x_ = None\n input_1 = torch.nn.functional.dropout(out, 0.1, True, False); out = None\n out_1 = torch.ops.test.baz(input_1); input_1 = None\n return (out_1,)\n \n# To see more debug info, please use `graph_module.print_readable()`", + "[qqiqyf6plgvsceqtyjt7qfdyz4oo3oixqvjcpsmypqbyyipyaif] example_inputs[0]: TensorMetadata(dtype=torch.float32, shape=torch.Size([1, 320, 128, 128]), stride=(5242880, 1, 40960, 320), device=device(type='cuda', index=0), layout=torch.strided, memory_format=torch.channels_last, storage_offset=0, storage_bytes=None, requires_grad=False, is_quantized=False, is_conj=False, is_neg=False, is_inference=False, is_sparse=False, is_coalesced=None, dense_dim=None, sparse_dim=None)", + "[xq2hdkbfkbcuye6rgtypayrkhqf4cntij2dsd24rei3lsknakkf] fx_kwargs[cudagraphs]: BoxedBool(value=False)", + "[du4vyrfyozrfxcf6kk6ma7oqwatapifazeelfsawmsiu6gjdtxp] deterministic_algorithms_settings: (False, False, True)", + "[qiptf2633zubseuei4bkisoq3not35l6lud6p23p4qmcsxiw2uq] cuda_matmul_settings: (False, True, True)", + "[svke3c6tlftklkb3z4oku47swr6tsobagmp2dlfuxz5mzlefnlw] torch_version: ", + "[poglqjwowp4gnkmehjby2lvdjrwuo5tbxa2gayd6smgasl2hgsd] system_info[device]: {'name': 'NVIDIA H100'}", + "[zhk6mbgyespwiq2kqql3qkec5aaj467ll4jlebwu35uhzivsy6u] system_info[version]: {'triton': '3.0.0+dedb7bdf33dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-835d4fc33500e1accafc5c5e00f4f73d87432c114860c04b68849bf6f942b8e5-dc767c8fadcf23ea82d79e257c37d44077eae7f681cf967565fd43e9c017937b-23d635e690d670bf61798e1259674b78c0ed5ba222ab6a455f329f27a758fc2d-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-20b017e9c4d858ab05e783f77df50b86c6d6eee5d79f3f4b158562b4a54f8443-f44338a31e0534290b08653050804c3fabbde403a6d3004ae04f0c28495f0802-e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855-a979896b9c0acfd41dd953b90bdc4b10968f7c0b45a286eae3f829aaddb2bb55-da771298f7bc45d24a61f35ef51742304421df1ab49d50bf1fc510dd5a46ea4b-dfa6e0edf1d43ef41636f628897e0ef501fd2b01159a3357f5c7130be523863b-71330f394e584b0df29595d49f6ac8ac0c5503db9147090dc58ad888cebac7be-f24adfd52383f7866791ebaa5d45a5d2cc826e56ee2fd285f438e85d201fe643-a34be0d3ae4b3ac9aede195cfda42f8a0a097b2bc9642fb59673ce6b3b607f10-36130a37af1b19a0dec569aa08d30b00c74c8f02b6b632999d86dea169146792-36d42f0429aae027cb985b53b9abc616fae4dad9e0ea03953e1e9fb46d0fb9a0-e5d2cb724c08d0ef4130f3ba858d22cf21f834bfd970a5388aa6ad2a6bab91f9', 'cuda': '12.2'}", + "[2qv36hmzaa3pkd42j2bzjmfpjvr32xx7ahdfiiepfwehvvcq45y] system_info[hash]: b2f391b9cfc799798db5e5f32606e9451cbd00c288b4ee846b49ad55396a924d", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[TYPE_CHECKING]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[enable_auto_functionalized_v2]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_progress]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[verbose_progress]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[fx_graph_cache]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[fx_graph_remote_cache]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bundle_triton_into_fx_graph_cache]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_local_cache]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[autotune_remote_cache]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[bundled_autotune_remote_cache]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_disable_caches]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[sleep_sec_TESTING_ONLY]: None", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[custom_op_default_layout_constraint]: needs_fixed_stride_order", + "[pikr7bbcoixfzftsazp5ggufhdklj24babfry77bl4nuvyrrcp4] inductor_config[triton_kernel_default_layout_constraint]: needs_fixed_stride_order", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp_wrapper]: False", + "[b4ha3ravs3qv237q65hpfqegbnoww7tf2ahcbu2i7xo6te5spqs] inductor_config[c_shim_version]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[dce]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[static_weight_shapes]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[size_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[nan_asserts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pick_loop_orders]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[inplace_buffers]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[allow_buffer_reuse]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[memory_planning]: False", + "[x75won4jmsgeb63pcvwr2y4eteyzzdhmf5rv6xhjppie4hx2yu5] inductor_config[memory_pool]: intermediates", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_harness]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[epilogue_fusion]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[epilogue_fusion_first]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[pattern_matcher]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[b2b_gemm_pass]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_pre_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[joint_custom_post_pass]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[pre_grad_custom_pass]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_cat_fx_passes]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[efficient_conv_bn_eval_fx_passes]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[is_predispatch]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[group_fusion]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[batch_fusion]: True", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[pre_grad_fusion_options]: {}", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[post_grad_fusion_options]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_locality]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[dynamic_scale_rblock]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_fuse_int_mm_with_mul]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[use_mixed_mm]: True", + "[zwmmbkdkarexuhbigurz5lfnhx64tht7fznecjkrvznh6rzivbv] inductor_config[fx_passes_numeric_check]: {'pre_grad': False, 'precision': 0.0001, 'num_iterations': 1, 'requires_optimizer': True}", + "[v2td5s4lnsvyxvaevy4chx6kc5h3mm2axazbgwimqule5zrzao7] inductor_config[mixed_mm_choice]: heuristic", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[reorder_for_compute_comm_overlap]: False", + "[ssupi7bu3rrhdpg2jyegzncu3kg3nnhklyliqvutaxgs7y7k3dx] inductor_config[reorder_for_compute_comm_overlap_passes]: ['reorder_compute_for_overlap', 'sink_waits', 'raise_comms']", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[reorder_for_peak_memory]: True", + "[lxxtoqhcoepwfokeiibd575gnxo3uzwiv4hmpomlwkpzqz3qzsh] inductor_config[estimate_op_runtime]: default", + "[yezuzjtg4h3jjur4jwtwiehbyixa7eonq4tqsqmwqve2lvvmrem] inductor_config[intra_node_bw]: 300", + "[5fxczt3ciyxitdhizb7sfsgn7fhpczcqsngttnt5ot2wyctk7co] inductor_config[inter_node_bw]: 25", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_pointwise]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[max_autotune_gemm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_same_precision]: False", + "[2y7luesktjrque3nr7qtxnum2mkbeegzdrsvkm3rvdlhqboajhx] inductor_config[max_autotune_gemm_backends]: ATEN,TRITON,CPP", + "[uqlsbif4zxd75vt522p52txyuguieipi2lwz5g5awt56lccqk7s] inductor_config[max_autotune_conv_backends]: ATEN,TRITON", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[max_autotune_gemm_search_space]: DEFAULT", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[autotune_fallback_to_aten]: True", + "[wft6ljqsfr3x4m7fa5zuyb7cwknky4irrxz4bjr6uzr2yiopxqj] inductor_config[unbacked_symint_fallback]: 8192", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[search_autotune_cache]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[save_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_in_subproc]: False", + "[iglov24t7x5ruci344aer2tm6nqshi4veuw4wxlssxtu46cx76m] inductor_config[max_autotune_subproc_result_timeout_seconds]: 60.0", + "[bh33ranllcgilhgmgr3qvygzxjm6isq5iexnfm3zx6fnr2zwlp2] inductor_config[max_autotune_subproc_graceful_timeout_seconds]: 1.0", + "[pwoh5aypf4fxbntdvwt67rppxorqos6xr3w7qzeun6kblbfg2ga] inductor_config[max_autotune_subproc_terminate_timeout_seconds]: 2.0", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[autotune_multi_device]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_tuning]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[coordinate_descent_check_all_directions]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[coordinate_descent_search_radius]: 1", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[autoheuristic_collect]: ", + "[jwbrgxes7vjqumngs5hyj6gn5nytv2whnppnzngvaagfmawhkkd] inductor_config[autoheuristic_use]: mixed_mm", + "[jvchmi66fvqzlemhr5fcqorz5trfdtdalzfagtj2aolmimwqhdq] inductor_config[autoheuristic_log_path]: DEFAULT", + "[4p2fdjlvxrcw7c7fvzm5huhtqxnro4kvkx56f7p5zyrxqkwooov] inductor_config[layout_opt_default]: 1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[layout_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_layout_optimization]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[keep_output_stride]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[warn_mix_layout]: False", + "[lkkae3meylaixfif4thncru4hjqeaislawjoghffrbwuscaagei] inductor_config[realize_reads_threshold]: 4", + "[rr5m5hsocoyodldz7vcvaizdwvm2rt34evmqdxvng7wz3tufvo6] inductor_config[realize_opcount_threshold]: 30", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[realize_acc_reads_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[fallback_random]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[implicit_fallbacks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aggressive_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_fusion]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[enabled_metric_tables]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[loop_ordering_after_fusion]: False", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[score_fusion_memory_threshold]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[benchmark_epilogue_fusion]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[max_epilogue_benchmarked_choices]: 1", + "[jykiys6ynafs3zdylwa5ggq6j655mxeh42d6mtdi22gffkrmiac] inductor_config[max_fusion_size]: 64", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[max_pointwise_cat_inputs]: 8", + "[yttmfmxblgcbsvbokguzowcorrcxz5uunxtcvsbe6nijgcx45he] inductor_config[unroll_reductions_threshold]: 8", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[comment_origin]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[conv_1x1_as_mm]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[split_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_kernel]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[constant_and_index_propagation]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[always_keep_tensor_constants]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[assert_indirect_indexing]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[compute_all_bounds]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernels]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[benchmark_combo_kernel]: False", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernels_autotune]: 1", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[combo_kernel_allow_mixed_sizes]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[combo_kernel_foreach_dynamic_shapes]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[joint_graph_constant_folding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[debug_index_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[emulate_precision_casts]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[is_nightly_or_source]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[developer_warnings]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[optimize_scatter_upon_const_tensor]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[global_cache_dir]: None", + "[j6c55jha5r2sdys2rwq7uqhtleea5dgjcye7nicfgft36v7xfvp] inductor_config[kernel_name_max_ops]: 10", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[shape_padding]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[comprehensive_padding]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_channels_last]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[disable_padding_cpu]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[padding_alignment_bytes]: 128", + "[dnnw5ks3yxrp7mwvihb2hh4tqx35ye637xt33x64kw4fvz2nyzg] inductor_config[padding_stride_threshold]: 1024", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[pad_outputs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[bw_outputs_user_visible]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[force_shape_pad]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[permute_fusion]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profiler_mark_wrapper_call]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[generate_intermediate_hooks]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[debug_ir_traceback]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth]: False", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[profile_bandwidth_regex]: ", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[profile_bandwidth_output]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[profile_bandwidth_with_do_bench_using_profiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[disable_cpp_codegen]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[freezing_discard_parameters]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[allow_stack_allocation]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[use_minimal_arrayref_interface]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[decompose_mem_bound_mm]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[assume_aligned_inputs]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[unsafe_ignore_unsupported_triton_autotune_args]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[check_stack_no_cycles_TESTING_ONLY]: False", + "[sz3im5ogc6asp7g4uqocnovype63tkdexzfrniv6hn2oank3biu] inductor_config[cpp.threads]: -1", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.no_redundant_loops]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.dynamic_threads]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.simdlen]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.min_chunk_size]: 1", + "[c7zj4qytmety6keurs3hsh5wn7foxp3dqx4kym2ucszzcb2ngrf] inductor_config[cpp.cxx]: (None, 'g++')", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_kernel_profile]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.weight_prepack]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_relu_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.inject_log1p_bug_TESTING_ONLY]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.vec_isa_ok]: None", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[cpp.descriptive_names]: original_aten", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[cpp.max_horizontal_fusion_size]: 16", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.fallback_scatter_reduce_sum]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_unsafe_math_opt_flag]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_floating_point_contract_flag]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_tiling_heuristics]: True", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cpp.gemm_max_k_slices]: 1", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_cache_blocking]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cpp.gemm_thread_factors]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[cpp.enable_loop_tail_vec]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cpp.enable_concat_linear]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_trees]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_skip_dynamic_graphs]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.slow_path_cudagraph_asserts]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cudagraph_trees_history_recording]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.cudagraph_support_input_mutation]: True", + "[ljdqgtysl3vdf7j6attlz5gmjg2ncihnveojfyubosplmkrjgra] inductor_config[triton.cudagraph_unexpected_rerecord_limit]: 128", + "[tuax46wac7rfv2trf5gcps6vleo3cq44lbnrdxtprvo3ljjaddj] inductor_config[triton.cudagraph_dynamic_shape_warn_limit]: 50", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraph_sync]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cudagraphs_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.fast_path_cudagraph_asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.skip_cudagraph_warmup]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_graph]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.debug_sync_kernel]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.dense_indexing]: False", + "[pr5nr4a7dthirgd2ljo3d2xakc63ywxugusu6mkmr6gmpeliyib] inductor_config[triton.max_tiles]: 2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.prefer_nd_tiling]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.autotune_pointwise]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.autotune_cublasLt]: True", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.autotune_at_compile_time]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_pointwise_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.tiling_prevents_reduction_fusion]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.unique_kernel_names]: True", + "[yrty22bseefglnysuoec4ji7j2rnaggdj3g33zzj7avogwfmgdw] inductor_config[triton.descriptive_names]: original_aten", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.persistent_reductions]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.cooperative_reductions]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.force_cooperative_reductions]: False", + "[vrl5ktomgtzox5xucd3np6vug3vyj6hwwzahqijuwpmamlv7ohi] inductor_config[triton.multi_kernel]: 0", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.divisible_by_16]: True", + "[fv6slhtedtydps5s5u2etitscliblzcidyitqf7krsv4e23fzk6] inductor_config[triton.min_split_scan_rblock]: 256", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.store_cubin]: False", + "[ebt2ncs4f5y7dn7btzi76mnouepvzad474tmp5iju4wiuumjl4s] inductor_config[triton.spill_threshold]: 16", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[triton.use_block_ptr]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[triton.inject_relu_bug_TESTING_ONLY]: None", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[triton.codegen_upcast_to_fp32]: True", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.output_path]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.debug_compile]: False", + "[ngkkx5e6z7erl6da23zb2cmsctz4yvaqyameyg5hbqln4wrhh7x] inductor_config[aot_inductor.debug_intermediate_value_printer]: 0", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[aot_inductor.filtered_kernel_names]: None", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_in_spec]: ", + "[v3hzzlv4tjgvp3pyhmzagjd25orl6n7nynoa7svlhhwk73b7u3c] inductor_config[aot_inductor.serialized_out_spec]: ", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.use_runtime_constant_folding]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.force_mmap_weights]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.package_cpp_only]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.metadata]: {}", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[aot_inductor.raise_error_on_ignored_optimization]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[aot_inductor.dump_aoti_minifier]: False", + "[4bryyl4ahh5whyg3zwqebpwmjnx6w77nqgqbdjlowju6lkqtn7w] inductor_config[aot_inductor.presets]: {}", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.arch]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.version]: None", + "[tvyftmtdmezlejo2xllu7awzv4pzc4vm4fub4b3gpl5jptjkosi] inductor_config[cuda.compile_opt_level]: -O1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_cuda_lto]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_ptxas_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.enable_debug_info]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.use_fast_math]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_max_profiling_configs]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cuda_cxx]: None", + "[aghvyrrgwvxijco2pk5wzc3cgmmthrbmgxitiibxuuscxdwrjd3] inductor_config[cuda.cutlass_backend_min_gemm_size]: 1", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[cuda.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[cuda.cutlass_op_allowlist_regex]: None", + "[lwkz5chtpji756gurqw4foijfi7zfgljtnn5nmnvdi2skpt4mgh] inductor_config[cuda.cutlass_op_denylist_regex]: pingpong", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[rocm.arch]: []", + "[oartxnko2l7d67tzwwm2otcumaut3n4wwcfgz3o377hmcveu5ft] inductor_config[rocm.ck_supported_arch]: ['gfx90a', 'gfx940', 'gfx941', 'gfx942']", + "[klfqjprnpfhcdurgvuikvc4rpd5ynkpk77toousr5h3u5roty6p] inductor_config[rocm.compile_opt_level]: -O2", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.is_debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.save_temps]: False", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.use_fast_math]: True", + "[cev5uo2jlwdhw2uyzcm7vr6cl23azjfw437f5r5lskm7spucos6] inductor_config[rocm.flush_denormals]: True", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.print_kernel_resource_usage]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.rocm_home]: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.ck_dir]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.generate_test_runner]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] inductor_config[rocm.n_max_profiling_configs]: None", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[rocm.use_preselected_instances]: False", + "[bsvfcwwoczx2rlkdz2eta6doujsymyihmi46hhwk6clrrvwcb6m] inductor_config[cpu_backend]: cpp", + "[caw4ly2z672k6kjfahoxwpajp5idhhtrpgf3ma2clylcp7c7aid] inductor_config[cuda_backend]: triton", + "[ljhgflgihidopsfsdcbqynv27nceykby3nutyd5jlcpq7n6e7l4] inductor_config[halide.cpu_target]: host", + "[wx7vmsmrdpk5ue2txlywp3lj3faqmdjphs5fgg2ehzsyno7uovg] inductor_config[halide.gpu_target]: host-cuda", + "[svgytlua5wcyeia7wq7e6zgh5tsueikrnzchmdmouvmkpfsc2zq] inductor_config[halide.scheduler_cuda]: Anderson2021", + "[k5ogk6345jvklsnu7g2njqstiz2g6pm5wmqpgg3kasrmuqwjvl6] inductor_config[halide.scheduler_cpu]: Adams2019", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.asserts]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.debug]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[halide.scan_kernels]: False", + "[h25wqx6vliw4j5rtzzbv6latydxyei3deyg6v7wzvnzryfktuki] inductor_config[external_matmul]: []", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.force_extern_kernel_in_multi_template]: False", + "[esstihe2nyydk4mhzpvox3qkajyu5y5t23hk3fi2me7jn75xi3o] inductor_config[test_configs.runtime_triton_dtype_assert]: False", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_pre_pass: None", + "[tquy2we2efmowuj4wuqzcfcfdcrkzkzmwdae6hprj7fa64jpusq] post_grad_custom_post_pass: None" + ], + "cache_bypass_reason": null, + "remote_cache_enabled": false, + "local_cache_enabled": true + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.913000 1500233 torch/_dynamo/guards.py:2315] {"dynamo_cpp_guards_str": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "0f04f735fae7ac445870b5956bc06ba9"} + + TREE_GUARD_MANAGER: + +- RootGuardManager + | +- DEFAULT_DEVICE: utils_device.CURRENT_DEVICE == None # _dynamo/output_graph.py:484 in init_ambient_guards + | +- GLOBAL_STATE: ___check_global_state() + | +- TORCH_FUNCTION_MODE_STACK: ___check_torch_function_mode_stack() + | +- GuardManager: source=L['x'], accessed_by=DictGetItemGuardAccessor('x') + | | +- TENSOR_MATCH: check_tensor(L['x'], Tensor, DispatchKeySet(CUDA, BackendSelect, ADInplaceOrView, AutogradCUDA), torch.float32, device=0, requires_grad=False, size=[1, 320, 128, 128], stride=[5242880, 1, 40960, 320]) + | | +- NO_HASATTR: hasattr(L['x'], '_dynamo_dynamic_indices') == False + | +- GuardManager: source=L['self'], accessed_by=DictGetItemGuardAccessor('self') + | | +- TYPE_MATCH: ___check_type_id(L['self'], 184117680) + | | +- GuardManager: source=L['self'].__dict__, accessed_by=GetGenericDictGuardAccessor + | | | +- GuardManager: source=L['self']._modules, accessed_by=DictGetItemGuardAccessor('_modules') + | | | | +- DICT_LENGTH: len(L['self']._modules) == 1 + | | | | +- GuardManager: source=L['self']._modules['in_layers'], accessed_by=DictGetItemGuardAccessor('in_layers') + | | | | | +- TYPE_MATCH: ___check_type_id(L['self']._modules['in_layers'], 113588064) + | | | | | +- LENGTH_CHECK: len(L['self']._modules['in_layers']) == 1 + | | | | | +- GuardManager: source=L['self']._modules['in_layers'].__dict__, accessed_by=GetGenericDictGuardAccessor + | | | | | | +- DICT_CONTAINS: not ___dict_contains('forward', L['self']._modules['in_layers'].__dict__) + | | | | | | +- DictGuardManager: source=L['self']._modules['in_layers']._modules, accessed_by=DictGetItemGuardAccessor('_modules') + | | | | | | | +- KeyValueManager pair at index=0 + | | | | | | | | +- KeyManager: GuardManager: source=list(L['self']._modules['in_layers']._modules.keys())[0] + | | | | | | | | | +- EQUALS_MATCH: list(L['self']._modules['in_layers']._modules.keys())[0] == '0' + | | | | | | | | +- ValueManager: GuardManager: source=L['self']._modules['in_layers']._modules['0'] + | | | | | | | | | +- TYPE_MATCH: ___check_type_id(L['self']._modules['in_layers']._modules['0'], 111068416) + | | | | | | | | | +- GuardManager: source=L['self']._modules['in_layers']._modules['0'].__dict__, accessed_by=GetGenericDictGuardAccessor + | | | | | | | | | | +- DICT_CONTAINS: not ___dict_contains('forward', L['self']._modules['in_layers']._modules['0'].__dict__) + | | | | | | | | | | +- GuardManager: source=L['self']._modules['in_layers']._modules['0'].p, accessed_by=DictGetItemGuardAccessor('p') + | | | | | | | | | | | +- EQUALS_MATCH: L['self']._modules['in_layers']._modules['0'].p == 0.1 + | | | | | | | | | | +- GuardManager: source=L['self']._modules['in_layers']._modules['0'].inplace, accessed_by=DictGetItemGuardAccessor('inplace') + | | | | | | | | | | | +- ID_MATCH: ___check_obj_id(L['self']._modules['in_layers']._modules['0'].inplace, 8907584) + | | | | | | | | | | +- GuardManager: source=L['self']._modules['in_layers']._modules['0'].training, accessed_by=DictGetItemGuardAccessor('training') + | | | | | | | | | | | +- ID_MATCH: ___check_obj_id(L['self']._modules['in_layers']._modules['0'].training, 8908032) + | | | +- GuardManager: source=L['self']._parameters, accessed_by=DictGetItemGuardAccessor('_parameters') + | | | | +- DICT_LENGTH: not L['self']._parameters + | +- GuardManager: source=G, accessed_by=GlobalsGuardAccessor + | | +- GuardManager: source=G['F'], accessed_by=DictGetItemGuardAccessor('F') + | | | +- ID_MATCH: ___check_obj_id(G['F'], 139720064954224) + | | | +- OBJECT_ALIASING: G['F'] is G['__import_torch_dot_nn_dot_modules_dot_dropout'].F + | | | +- GuardManager: source=G['F'].gelu, accessed_by=GetAttrGuardAccessor(gelu) + | | | | +- ID_MATCH: ___check_obj_id(G['F'].gelu, 139721962700240) + | | | +- GuardManager: source=G['F'].dropout, accessed_by=GetAttrGuardAccessor(dropout) + | | | | +- ID_MATCH: ___check_obj_id(G['F'].dropout, 139720032587424) + | | +- GuardManager: source=G['torch'], accessed_by=DictGetItemGuardAccessor('torch') + | | | +- ID_MATCH: ___check_obj_id(G['torch'], 139724697307792) + | | | +- GuardManager: source=G['torch'].ops, accessed_by=GetAttrGuardAccessor(ops) + | | | | +- ID_MATCH: ___check_obj_id(G['torch'].ops, 139720033503648) + | | | | +- GuardManager: source=G['torch'].ops.test, accessed_by=GetAttrGuardAccessor(test) + | | | | | +- ID_MATCH: ___check_obj_id(G['torch'].ops.test, 139719717837520) + | | | | | +- GuardManager: source=G['torch'].ops.test.baz, accessed_by=GetAttrGuardAccessor(baz) + | | | | | | +- ID_MATCH: ___check_obj_id(G['torch'].ops.test.baz, 139710547383760) + | | +- GuardManager: source=G['__builtins_dict___0'], accessed_by=DictGetItemGuardAccessor('__builtins_dict___0') + | | | +- GuardManager: source=G['__builtins_dict___0']['iter'], accessed_by=DictGetItemGuardAccessor('iter') + | | | | +- ID_MATCH: ___check_obj_id(G['__builtins_dict___0']['iter'], 139724706059008) + | | +- GuardManager: source=G['__import_torch_dot_nn_dot_modules_dot_module'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_nn_dot_modules_dot_module') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_modules_dot_module'], 139721387922336) + | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_modules_dot_module']._global_forward_hooks, accessed_by=GetAttrGuardAccessor(_global_forward_hooks) + | | | | +- DICT_LENGTH: not G['__import_torch_dot_nn_dot_modules_dot_module']._global_forward_hooks + | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_modules_dot_module']._global_backward_hooks, accessed_by=GetAttrGuardAccessor(_global_backward_hooks) + | | | | +- DICT_LENGTH: not G['__import_torch_dot_nn_dot_modules_dot_module']._global_backward_hooks + | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_modules_dot_module']._global_forward_pre_hooks, accessed_by=GetAttrGuardAccessor(_global_forward_pre_hooks) + | | | | +- DICT_LENGTH: not G['__import_torch_dot_nn_dot_modules_dot_module']._global_forward_pre_hooks + | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_modules_dot_module']._global_backward_pre_hooks, accessed_by=GetAttrGuardAccessor(_global_backward_pre_hooks) + | | | | +- DICT_LENGTH: not G['__import_torch_dot_nn_dot_modules_dot_module']._global_backward_pre_hooks + | | +- GuardManager: source=G['__import_torch_dot_nn_dot_modules_dot_dropout'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_nn_dot_modules_dot_dropout') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_modules_dot_dropout'], 139720031577856) + | | | +- GuardManager: source=G['__import_torch_dot_nn_dot_modules_dot_dropout'].F, accessed_by=GetAttrGuardAccessor(F) + | | | | +- OBJECT_ALIASING: G['F'] is G['__import_torch_dot_nn_dot_modules_dot_dropout'].F + | | +- GuardManager: source=G['__import_torch_dot_nn_dot_modules_dot_container'], accessed_by=DictGetItemGuardAccessor('__import_torch_dot_nn_dot_modules_dot_container') + | | | +- ID_MATCH: ___check_obj_id(G['__import_torch_dot_nn_dot_modules_dot_container'], 139720031411696) + +V1206 15:18:21.914000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "2967ad97a7a8f2d0901c376ff0ef093c"} + { + "name": "entire_frame_compile", + "ts": 1733527101913985.8, + "args": { + "fn_name": "_compile.compile_inner", + "compile_id": "0/0" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } +V1206 15:18:21.916000 1500233 torch/_dynamo/utils.py:1045] {"compilation_metrics": {"pre_grad_pass_time_us": 11924, "joint_graph_pass_time_us": 89462, "feature_usage": {"pytorch/remote_cache:fx_graph_memcache_version": true, "pytorch/remote_cache:bundle_triton_into_fx_graph_cache_v2": true}, "post_grad_pass_time_us": 17215, "inductor_code_gen_cumulative_compile_time_us": 529174, "inductor_cumulative_compile_time_us": 706575, "aot_autograd_cumulative_compile_time_us": 1307090, "dynamo_cumulative_compile_time_us": 1659332, "frame_key": "1", "co_name": "forward", "co_filename": "/data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py", "co_firstlineno": 11123, "cache_size": 0, "accumulated_cache_size": 0, "guard_count": 49, "shape_env_guard_count": 0, "graph_op_count": 3, "graph_node_count": 5, "graph_input_count": 1, "fail_type": null, "fail_reason": null, "fail_user_frame_filename": null, "fail_user_frame_lineno": null, "non_compliant_ops": ["test::baz"], "compliant_custom_ops": [], "restart_reasons": [], "dynamo_time_before_restart_s": 0.0, "has_guarded_code": true, "config_suppress_errors": false, "config_inline_inbuilt_nn_modules": true, "specialize_float": false, "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._refs\", \"torch.testing\", \"torch._prims\", \"torch.distributions\", \"torch._decomp\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": false, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}", "is_forward": true, "dynamo_compile_time_before_restart_us": 0}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0} +V1206 15:18:21.916000 1500233 torch/_dynamo/utils.py:1288] {"chromium_event": {}, "compiled_autograd_id": null, "frame_id": 0, "frame_compile_id": 0, "attempt": 0, "has_payload": "188d1c02e0f0d39aa82bec9ce598f130"} + { + "name": "dynamo", + "ts": 1733527101916886.5, + "args": { + "compile_id": "0/0", + "frame_key": "1", + "co_name": "forward", + "co_filename": "/data/users/xmfan/a/pytorch/test/inductor/test_torchinductor.py", + "co_firstlineno": 11123, + "cache_size": 0, + "accumulated_cache_size": 0, + "guard_count": 49, + "shape_env_guard_count": 0, + "graph_op_count": 3, + "graph_node_count": 5, + "graph_input_count": 1, + "fail_type": null, + "fail_reason": null, + "fail_user_frame_filename": null, + "fail_user_frame_lineno": null, + "non_compliant_ops": [ + "test::baz" + ], + "compliant_custom_ops": [], + "restart_reasons": [], + "dynamo_time_before_restart_s": 0.0, + "has_guarded_code": true, + "dynamo_config": "{\"verify_correctness\": false, \"minimum_call_count\": 1, \"dead_code_elimination\": true, \"cache_size_limit\": 8, \"accumulated_cache_size_limit\": 256, \"skip_code_recursive_on_cache_limit_hit\": true, \"fail_on_cache_limit_hit\": false, \"specialize_int\": false, \"specialize_float\": false, \"dynamic_shapes\": true, \"use_lazy_graph_module\": true, \"assume_static_by_default\": true, \"automatic_dynamic_shapes\": true, \"force_parameter_static_shapes\": true, \"force_nn_module_property_static_shapes\": true, \"allow_ignore_mark_dynamic\": false, \"guard_nn_modules\": true, \"guard_nn_modules_using_dict_tags\": true, \"suppress_errors\": false, \"replay_record_enabled\": false, \"rewrite_assert_with_torch_assert\": true, \"disable\": false, \"cprofile\": false, \"skipfiles_inline_module_allowlist\": {}, \"allowed_functions_module_string_ignorelist\": [\"torch._refs\", \"torch.testing\", \"torch._prims\", \"torch.distributions\", \"torch._decomp\"], \"capture_scalar_outputs\": false, \"capture_dynamic_output_shape_ops\": false, \"prefer_deferred_runtime_asserts_over_guards\": false, \"allow_complex_guards_as_runtime_asserts\": false, \"force_unspec_int_unbacked_size_like_on_torchrec_kjt\": false, \"optimize_ddp\": true, \"do_not_emit_runtime_asserts\": false, \"_ddp_optimization_mode\": [\"ddp_optimizer\", \"python_reducer\", \"python_reducer_without_compiled_forward\", \"no_optimization\"], \"skip_torchrec\": true, \"optimize_ddp_lazy_compile\": false, \"skip_fsdp_guards\": true, \"skip_fsdp_hooks\": true, \"skip_nnmodule_hook_guards\": true, \"skip_no_tensor_aliasing_guards_on_parameters\": true, \"raise_on_ctx_manager_usage\": true, \"raise_on_unsafe_aot_autograd\": false, \"error_on_nested_jit_trace\": true, \"error_on_nested_fx_trace\": true, \"allow_rnn\": false, \"capture_sparse_compute\": true, \"error_on_recompile\": false, \"report_guard_failures\": true, \"trace_numpy\": true, \"numpy_default_float\": \"float64\", \"numpy_default_complex\": \"complex128\", \"numpy_default_int\": \"int64\", \"use_numpy_random_stream\": false, \"enable_cpp_guard_manager\": true, \"inline_inbuilt_nn_modules\": true, \"record_compile_time_instruction_count\": false, \"cudagraph_backend_keep_input_mutation\": false, \"cudagraph_backend_support_input_mutation\": false, \"only_allow_pt2_compliant_ops\": false, \"capture_autograd_function\": true, \"capture_func_transforms\": true, \"fake_tensor_cache_enabled\": true, \"fake_tensor_cache_crosscheck_enabled\": false, \"compiled_autograd\": false, \"compiled_autograd_kwargs_override\": {}, \"enable_compiler_collectives\": false, \"automatic_dynamic_local_pgo\": false, \"automatic_dynamic_remote_pgo\": null}" + }, + "ph": "E", + "cat": "dynamo_timed", + "tid": 0, + "pid": 0 + } \ No newline at end of file diff --git a/tests/integration_test.rs b/tests/integration_test.rs index 345226b..87cc9cd 100644 --- a/tests/integration_test.rs +++ b/tests/integration_test.rs @@ -17,7 +17,6 @@ fn test_parse_simple() { "failures_and_restarts.html", "0_0_0/inductor_post_grad_graph", "0_0_0/inductor_output_code", - "0_0_0/dynamo_guards", ]; // Read the test file // simple.log was generated from the following: @@ -44,20 +43,17 @@ fn test_parse_simple() { fn test_parse_compilation_metrics() { let expected_files = [ "0_0_1/dynamo_output_graph", - "0_0_1/dynamo_guards", "0_0_1/compilation_metrics", "1_0_1/dynamo_output_graph", - "1_0_1/dynamo_guards", "1_0_1/compilation_metrics", "2_0_0/dynamo_output_graph", - "2_0_0/dynamo_guards", "2_0_0/compilation_metrics", "index.html", "failures_and_restarts.html", ]; // Read the test file - // simple.log was generated from the following: - // TORCH_TRACE=~/trace_logs/test python test/inductor/test_torchinductor.py -k TORCH_TRACE=~/trace_logs/comp_metrics python test/dynamo/test_misc.py -k test_graph_break_compilation_metrics + // comp_metrics.log was generated from the following: + // TORCH_TRACE=~/trace_logs/comp_metrics python test/dynamo/test_misc.py -k test_graph_break_compilation_metrics let path = Path::new("tests/inputs/comp_metrics.log").to_path_buf(); let config = tlparse::ParseConfig { strict: true, @@ -85,8 +81,8 @@ fn test_parse_compilation_failures() { "failures_and_restarts.html", ]; // Read the test file - // simple.log was generated from the following: - // TORCH_TRACE=~/trace_logs/test python test/inductor/test_torchinductor.py -k TORCH_TRACE=~/trace_logs/comp_metrics python test/dynamo/test_misc.py -k test_graph_break_compilation_metrics_on_failure + // comp_failure.log was generated from the following: + // TORCH_TRACE=~/trace_logs/comp_metrics python test/dynamo/test_misc.py -k test_graph_break_compilation_metrics_on_failure let path = Path::new("tests/inputs/comp_failure.log").to_path_buf(); let config = tlparse::ParseConfig { strict: true, @@ -109,7 +105,8 @@ fn test_parse_compilation_failures() { fn test_parse_artifact() { let expected_files = ["0_0_0/fx_graph_cache_hash", "index.html"]; // Read the test file - // simple.log was generated from the following: + // artifacts.log was generated from the following: + // NOTE: this test command looks wrong, and is not producing anything close to artifacts.log // TORCH_TRACE=~/trace_logs/test python test/inductor/test_torchinductor.py -k TORCH_TRACE=~/trace_logs/comp_metrics python test/dynamo/test_misc.py -k test_graph_break_compilation_metrics_on_failure let path = Path::new("tests/inputs/artifacts.log").to_path_buf(); let config = tlparse::ParseConfig { @@ -133,8 +130,8 @@ fn test_parse_artifact() { fn test_parse_chromium_event() { let expected_files = ["chromium_events.json", "index.html"]; // Read the test file - // simple.log was generated from the following: - // TORCH_TRACE=~/trace_logs/test python test/inductor/test_torchinductor.py -k TORCH_TRACE=~/trace_logs/comp_metrics python test/dynamo/test_misc.py -k test_graph_break_compilation_metrics_on_failure + // chromium_events.log was generated from the following: + // TORCH_TRACE=~/trace_logs/comp_metrics python test/dynamo/test_misc.py -k test_graph_break_compilation_metrics_on_failure let path = Path::new("tests/inputs/chromium_events.log").to_path_buf(); let config = tlparse::ParseConfig { strict: true, @@ -156,8 +153,9 @@ fn test_parse_chromium_event() { #[test] fn test_cache_hit_miss() { let expected_files = [ - "1_0_0/fx_graph_cache_miss_8", - "1_0_0/fx_graph_cache_hit_17", + "1_0_0/fx_graph_cache_miss_33.json", + "1_0_0/fx_graph_cache_miss_9.json", + "1_0_0/fx_graph_cache_hit_20.json", "index.html", ]; // Generated via TORCH_TRACE=~/trace_logs/test python test/inductor/test_codecache.py -k test_flex_attention_caching