#8865: Port ttnn changed ops to dispatch profiling infra (#10772)

* #8865: Port ttnn changed ttnn ops to dispatch profiling infra
* #8865: Update reference times for dispatch profiling infra

tests/ttnn/profiling/ops_for_profiling.py

@@ -82,6 +82,16 @@ def bcast_hw_shape_func_11(input_shape):
     return input_shape, input_shape_1
 
 
+def bcast_h_shape_func_1(input_shape):
+    input_shape_1 = [input_shape[-4], input_shape[-3], 1, input_shape[-1]]
+    return input_shape, input_shape_1
+
+
+def bcast_w_shape_func_1(input_shape):
+    input_shape_1 = [input_shape[-4], input_shape[-3], input_shape[-2], 1]
+    return input_shape, input_shape_1
+
+
 def complex_add(x, y):
     tt_lib.tensor.complex_add(
         x, y, tt_lib.tensor.MemoryConfig(tt_lib.tensor.TensorMemoryLayout.INTERLEAVED, tt_lib.tensor.BufferType.DRAM)
@@ -151,15 +161,15 @@ def unary_pow_bw(x, y):
 
 
 def clamp_bw(x, y):
-    ttnn.clamp_bw(x, y, 0.1, 0.9)
+    ttnn.clamp_bw(x, y, min=0.1, max=0.9)
 
 
 def clamp_min_bw(x, y):
-    ttnn.clamp_min_bw(x, y, 0.1)
+    ttnn.clamp_bw(x, y, min=0.1)
 
 
 def clamp_max_bw(x, y):
-    ttnn.clamp_max_bw(x, y, 0.9)
+    ttnn.clamp_bw(x, y, max=0.9)
 
 
 def gelu_bw_none(x, y):
@@ -207,7 +217,7 @@ def unary_eq_bw(x, y):
 
 
 def logiteps_bw(x, y):
-    ttnn.logiteps_bw(x, y, 0.0001)
+    ttnn.logiteps_bw(x, y, eps=0.0001)
 
 
 def fmod_bw(x, y):
@@ -418,23 +428,23 @@ def angle_bw(x, y):
 
 
 def celu_bw(x, y):
-    ttnn.celu_bw(x, y, 1)
+    ttnn.celu_bw(x, y, alpha=1)
 
 
 def hardshrink_bw(x, y):
-    ttnn.hardshrink_bw(x, y, 0.5)
+    ttnn.hardshrink_bw(x, y, lambd=0.5)
 
 
 def leaky_relu_bw(x, y):
-    ttnn.leaky_relu_bw(x, y, 0.3)
+    ttnn.leaky_relu_bw(x, y, negative_slope=0.3)
 
 
 def softshrink_bw(x, y):
-    ttnn.softshrink_bw(x, y, 0.5)
+    ttnn.softshrink_bw(x, y, lambd=0.5)
 
 
 def unary_div_bw(x, y):
-    ttnn.div_bw(x, y, 3, round_mode="None")
+    ttnn.div_bw(x, y, 3.0, round_mode="None")
 
 
 all_binary_ops = [
@@ -450,6 +460,16 @@ def unary_div_bw(x, y):
         "op": ttnn.mul,
         "name": "ttnn.mul",
     },
+    {
+        "op": ttnn.mul,
+        "name": "ttnn.mul_bcast_h",
+        "shape_func": bcast_h_shape_func_1,
+    },
+    {
+        "op": ttnn.mul,
+        "name": "ttnn.mul_bcast_w",
+        "shape_func": bcast_w_shape_func_1,
+    },
     {
         "op": ttnn.mul,
         "name": "ttnn.mul_bcast_hw",
@@ -646,7 +666,7 @@ def unary_div_bw(x, y):
     },
     {
         "op": ttnn.embedding,
-        "name": "tt_lib.tensor.embeddings",
+        "name": "ttnn.embedding",
         "layout": "ROW_MAJOR",
         "shape_func": embeddings_shape_func,
     },
@@ -1170,11 +1190,11 @@ def leaky_relu(x):
 
 
 def softshrink(x):
-    ttnn.softshrink(x, 70)
+    ttnn.softshrink(x, lambd=70)
 
 
 def hardshrink(x):
-    ttnn.hardshrink(x, 1)
+    ttnn.hardshrink(x, lambd=1)
 
 
 def elu(x):
@@ -1194,7 +1214,7 @@ def bias_gelu_unary(x):
 
 
 def logit(x):
-    ttnn.logit(x, 0.0001)
+    ttnn.logit(x, eps=0.0001)
 
 
 def logical_andi(x):
@@ -1309,14 +1329,6 @@ def empty(x):
     ttnn.empty(shape=x.get_legacy_shape(), dtype=x.get_dtype(), layout=x.get_layout(), device=x.device())
 
 
-def tril(x):
-    ttnn.tril(x, 1)
-
-
-def triu(x):
-    ttnn.triu(x, 1)
-
-
 def sum_dim_2(x):
     ttnn.sum(x, dim=2)
 
@@ -1951,6 +1963,7 @@ def clone(x):
     {
         "op": tilize,
         "name": "ttnn.tilize",
+        "layout": "ROW_MAJOR",
     },
     {
         "op": tt_lib.tensor.untilize,
@@ -1968,6 +1981,7 @@ def clone(x):
     {
         "op": ttnn.tilize_with_zero_padding,
         "name": "ttnn.tilize_with_zero_padding",
+        "layout": "ROW_MAJOR",
     },
     {
         "op": pad,
@@ -2022,12 +2036,12 @@ def clone(x):
         "name": "ttnn.empty",
     },
     {
-        "op": tril,
+        "op": ttnn.tril,
         "name": "ttnn.tril",
         "num_repeats": 3,
     },
     {
-        "op": triu,
+        "op": ttnn.triu,
         "name": "ttnn.triu",
         "num_repeats": 3,
     },
@@ -2181,16 +2195,16 @@ def clone(x):
         "name": "tt_lib.tensor.mean_hw",
     },
     {
-        "op": tt_lib.tensor.var_hw,
-        "name": "tt_lib.tensor.var_hw",
+        "op": ttnn.var_hw,
+        "name": "ttnn.var_hw",
     },
     {
         "op": logical_noti,
         "name": "tt_lib.tensor.logical_noti",
     },
     {
-        "op": tt_lib.tensor.std_hw,
-        "name": "tt_lib.tensor.std_hw",
+        "op": ttnn.std_hw,
+        "name": "ttnn.std_hw",
     },
     {
         "op": ttnn.normalize_hw,
@@ -2534,18 +2548,23 @@ def div_bw(x, y, z):
     ttnn.div_bw(x, y, z, round_mode="None")
 
 
+def add_bw(x, y, z):
+    ttnn.add_bw(x, y, z)
+
+
 def primary_moreh_norm_backward(x, y, z):
     tt_lib.operations.primary.moreh_norm_backward(x, y, z, p=2.0)
 
 
-def fused_linear(x, weight, bias):
+def linear(x, weight, bias):
     ttnn.linear(x, weight, bias=bias)
 
 
-def fused_linear_shape_func(input_shape):
-    x_shape = [1, 1, input_shape[-2], input_shape[-1]]
-    weight_shape = [1, 1, input_shape[-2], input_shape[-1]]
-    bias_shape = [1, 1, 32, input_shape[-1]]
+def linear_shape_func(input_shape):
+    N = input_shape[-1]
+    x_shape = [1, input_shape[-2], N]
+    weight_shape = [N, N]
+    bias_shape = [1, N]
     return x_shape, weight_shape, bias_shape
 
 
@@ -2634,7 +2653,7 @@ def fused_linear_shape_func(input_shape):
         "name": "ttnn.min_bw",
     },
     {
-        "op": ttnn.add_bw,
+        "op": add_bw,
         "name": "ttnn.add_bw",
     },
     # {
@@ -2726,9 +2745,9 @@ def fused_linear_shape_func(input_shape):
         "name": "tt_lib.tensor.moreh_norm_backward",
     },
     {
-        "op": fused_linear,
+        "op": linear,
         "name": "ttnn.linear",
-        "shape_func": fused_linear_shape_func,
+        "shape_func": linear_shape_func,
     },
     {
         "op": ttnn.ge_bw,