Skip to content

Commit

Permalink
Update img, add assests
Browse files Browse the repository at this point in the history
  • Loading branch information
@ravi9
ravi9 committed Nov 23, 2024
1 parent 28e1b53 commit b59670b
Show file tree
Hide file tree
Showing 15 changed files with 304 additions and 0 deletions.
54 changes: 54 additions & 0 deletions ...llm_diffusion_serving_app/assets/benchmark_results_20241123_044407/benchmark_results.json
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,54 @@
{
"timestamp": "2024-11-23T04:44:07.510110",
"hardware_config": {
"cpu_model": "Intel(R) Xeon(R) Platinum 8488C",
"cpu_count": "64",
"threads_per_core": "2",
"cores_per_socket": "32",
"socket_count": "1",
"total_memory": "247.71 GB"
},
"software_versions": {
"Python": "3.9.20",
"TorchServe": "0.12.0",
"OpenVINO": "2024.5.0",
"PyTorch": "2.5.1+cpu",
"Transformers": "4.46.3",
"Diffusers": "0.31.0"
},
"benchmark_results": [
{
"run_mode": "eager",
"warmup_time": 11.164182662963867,
"statistics": {
"mean": 10.437215328216553,
"std": 0.0,
"all_iterations": [
10.437215328216553
]
}
},
{
"run_mode": "tc_inductor",
"warmup_time": 83.48197150230408,
"statistics": {
"mean": 8.774884462356567,
"std": 0.0,
"all_iterations": [
8.774884462356567
]
}
},
{
"run_mode": "tc_openvino",
"warmup_time": 52.01788377761841,
"statistics": {
"mean": 2.633979082107544,
"std": 0.0,
"all_iterations": [
2.633979082107544
]
}
}
]
}
Binary file added ...sion_serving_app/assets/benchmark_results_20241123_044407/image-eager-final.png
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added ...erving_app/assets/benchmark_results_20241123_044407/image-tc_inductor-final.png
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added ...erving_app/assets/benchmark_results_20241123_044407/image-tc_openvino-final.png
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
54 changes: 54 additions & 0 deletions ...ases/llm_diffusion_serving_app/assets/benchmark_results_20241123_044407/profile-eager.txt
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,54 @@
CPU Time Profile (sort_by='cpu_time_total', row_limit=20):
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
Name Self CPU % Self CPU CPU total % CPU total CPU time avg CPU Mem Self CPU Mem # of Calls
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
model_inference 7.36% 768.176ms 100.00% 10.437s 10.437s 159.56 Mb -34.34 Gb 1
aten::linear 0.17% 17.396ms 42.24% 4.409s 1.359ms 6.96 Gb 0 b 3245
aten::conv2d 0.01% 1.073ms 36.07% 3.765s 15.689ms 5.08 Gb 0 b 240
aten::convolution 0.02% 1.635ms 36.06% 3.764s 15.684ms 5.08 Gb 0 b 240
aten::_convolution 0.02% 2.382ms 36.05% 3.763s 15.677ms 5.08 Gb 0 b 240
aten::mkldnn_convolution 35.62% 3.718s 36.03% 3.760s 15.667ms 5.08 Gb -695.04 Mb 240
aten::addmm 29.53% 3.082s 29.94% 3.125s 2.065ms 5.00 Gb 5.00 Gb 1513
aten::matmul 0.12% 12.424ms 11.90% 1.242s 717.148us 1.97 Gb 0 b 1732
aten::mm 11.70% 1.221s 11.71% 1.222s 705.429us 1.97 Gb 1.97 Gb 1732
aten::scaled_dot_product_attention 0.04% 4.173ms 6.48% 675.892ms 1.117ms 916.37 Mb -28.39 Mb 605
aten::_scaled_dot_product_flash_attention_for_cpu 6.28% 655.775ms 6.44% 671.718ms 1.110ms 944.76 Mb -3.17 Gb 605
aten::group_norm 0.01% 1.539ms 1.91% 199.132ms 930.523us 4.45 Gb -63.03 Mb 214
aten::native_group_norm 1.84% 191.593ms 1.85% 193.122ms 902.437us 4.45 Gb 64 b 214
aten::silu 1.58% 164.591ms 1.58% 164.591ms 671.799us 4.36 Gb 4.36 Gb 245
aten::copy_ 1.14% 119.234ms 1.14% 119.234ms 39.144us 0 b 0 b 3046
aten::add 1.12% 116.458ms 1.12% 116.488ms 101.206us 3.39 Gb 3.39 Gb 1151
aten::div 0.90% 93.900ms 0.99% 103.540ms 156.405us 2.64 Gb 2.64 Gb 662
aten::layer_norm 0.03% 3.088ms 0.67% 69.474ms 74.704us 1.33 Gb -2.66 Mb 930
aten::native_layer_norm 0.57% 59.530ms 0.64% 66.387ms 71.384us 1.34 Gb 11.55 Kb 930
aten::upsample_nearest2d 0.35% 36.769ms 0.56% 58.545ms 5.322ms 1003.50 Mb 903.34 Mb 11
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
Self CPU time total: 10.437s


Memory Usage Profile (sort_by='cpu_memory_usage', row_limit=20):
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
Name Self CPU % Self CPU CPU total % CPU total CPU time avg CPU Mem Self CPU Mem # of Calls
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
aten::empty 0.13% 14.060ms 0.13% 14.060ms 2.320us 14.99 Gb 14.99 Gb 6061
aten::linear 0.17% 17.396ms 42.24% 4.409s 1.359ms 6.96 Gb 0 b 3245
aten::conv2d 0.01% 1.073ms 36.07% 3.765s 15.689ms 5.08 Gb 0 b 240
aten::convolution 0.02% 1.635ms 36.06% 3.764s 15.684ms 5.08 Gb 0 b 240
aten::_convolution 0.02% 2.382ms 36.05% 3.763s 15.677ms 5.08 Gb 0 b 240
aten::mkldnn_convolution 35.62% 3.718s 36.03% 3.760s 15.667ms 5.08 Gb -695.04 Mb 240
aten::addmm 29.53% 3.082s 29.94% 3.125s 2.065ms 5.00 Gb 5.00 Gb 1513
aten::native_group_norm 1.84% 191.593ms 1.85% 193.122ms 902.437us 4.45 Gb 64 b 214
aten::group_norm 0.01% 1.539ms 1.91% 199.132ms 930.523us 4.45 Gb -63.03 Mb 214
aten::silu 1.58% 164.591ms 1.58% 164.591ms 671.799us 4.36 Gb 4.36 Gb 245
aten::add 1.12% 116.458ms 1.12% 116.488ms 101.206us 3.39 Gb 3.39 Gb 1151
aten::div 0.90% 93.900ms 0.99% 103.540ms 156.405us 2.64 Gb 2.64 Gb 662
aten::matmul 0.12% 12.424ms 11.90% 1.242s 717.148us 1.97 Gb 0 b 1732
aten::mm 11.70% 1.221s 11.71% 1.222s 705.429us 1.97 Gb 1.97 Gb 1732
aten::empty_like 0.02% 2.222ms 0.06% 5.847ms 6.674us 1.82 Gb 0 b 876
aten::gelu 0.39% 40.409ms 0.39% 40.409ms 129.516us 1.78 Gb 1.78 Gb 312
aten::mul 0.14% 14.857ms 0.15% 15.175ms 42.387us 1.77 Gb 1.77 Gb 358
aten::empty_strided 0.05% 5.595ms 0.05% 5.595ms 3.011us 1.38 Gb 1.38 Gb 1858
aten::native_layer_norm 0.57% 59.530ms 0.64% 66.387ms 71.384us 1.34 Gb 11.55 Kb 930
aten::layer_norm 0.03% 3.088ms 0.67% 69.474ms 74.704us 1.33 Gb -2.66 Mb 930
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
Self CPU time total: 10.437s
54 changes: 54 additions & 0 deletions ...lm_diffusion_serving_app/assets/benchmark_results_20241123_044407/profile-tc_inductor.txt
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,54 @@
CPU Time Profile (sort_by='cpu_time_total', row_limit=20):
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
Name Self CPU % Self CPU CPU total % CPU total CPU time avg CPU Mem Self CPU Mem # of Calls
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
Torch-Compiled Region 11.15% 978.694ms 243.17% 21.338s 1.778s 49.11 Mb -5.99 Gb 12
model_inference 0.31% 27.435ms 100.00% 8.775s 8.775s 159.56 Mb -467.16 Mb 1
aten::addmm 33.55% 2.944s 34.48% 3.026s 1.995ms 54.14 Mb 54.14 Mb 1517
aten::convolution 0.03% 2.311ms 32.58% 2.859s 11.911ms 5.08 Gb 0 b 240
aten::_convolution 0.03% 2.694ms 32.55% 2.856s 11.901ms 5.08 Gb 0 b 240
aten::mkldnn_convolution 32.49% 2.851s 32.52% 2.854s 11.890ms 5.08 Gb 0 b 240
aten::mm 13.34% 1.170s 13.35% 1.171s 677.765us 2.50 Kb 2.50 Kb 1728
aten::_scaled_dot_product_flash_attention_for_cpu 7.41% 650.199ms 7.65% 671.659ms 1.110ms 944.76 Mb -3.17 Gb 605
aten::linear 0.01% 1.257ms 1.15% 101.083ms 523.748us 54.14 Mb 0 b 193
aten::copy_ 1.00% 87.585ms 1.00% 87.585ms 42.190us 0 b 0 b 2076
aten::to 0.00% 408.486us 0.15% 13.382ms 21.377us 478.68 Mb 0 b 626
aten::_to_copy 0.01% 1.300ms 0.15% 12.973ms 23.376us 478.68 Mb 1.25 Kb 555
aten::transpose 0.08% 6.832ms 0.12% 10.230ms 2.589us 0 b 0 b 3951
TorchDynamo Cache Lookup 0.09% 8.074ms 0.09% 8.074ms 672.829us 0 b 0 b 12
aten::empty 0.08% 6.633ms 0.08% 6.633ms 2.915us 3.21 Gb 3.21 Gb 2276
aten::empty_like 0.03% 2.893ms 0.07% 6.463ms 10.682us 916.37 Mb 0 b 605
aten::scaled_dot_product_attention 0.00% 182.884us 0.07% 6.327ms 197.709us 6.02 Mb -192.50 Kb 32
aten::expand 0.04% 3.462ms 0.06% 4.878ms 3.115us 0 b 0 b 1566
aten::as_strided 0.06% 4.864ms 0.06% 4.864ms 0.875us 0 b 0 b 5561
aten::gelu 0.05% 4.537ms 0.05% 4.537ms 141.789us 24.06 Mb 24.06 Mb 32
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
Self CPU time total: 8.775s


Memory Usage Profile (sort_by='cpu_memory_usage', row_limit=20):
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
Name Self CPU % Self CPU CPU total % CPU total CPU time avg CPU Mem Self CPU Mem # of Calls
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
aten::resize_ 0.02% 1.956ms 0.02% 1.956ms 8.050us 5.08 Gb 5.08 Gb 243
aten::convolution 0.03% 2.311ms 32.58% 2.859s 11.911ms 5.08 Gb 0 b 240
aten::_convolution 0.03% 2.694ms 32.55% 2.856s 11.901ms 5.08 Gb 0 b 240
aten::mkldnn_convolution 32.49% 2.851s 32.52% 2.854s 11.890ms 5.08 Gb 0 b 240
aten::empty 0.08% 6.633ms 0.08% 6.633ms 2.915us 3.21 Gb 3.21 Gb 2276
aten::empty_strided 0.05% 4.505ms 0.05% 4.505ms 3.883us 1.36 Gb 1.36 Gb 1160
aten::_scaled_dot_product_flash_attention_for_cpu 7.41% 650.199ms 7.65% 671.659ms 1.110ms 944.76 Mb -3.17 Gb 605
aten::empty_like 0.03% 2.893ms 0.07% 6.463ms 10.682us 916.37 Mb 0 b 605
aten::to 0.00% 408.486us 0.15% 13.382ms 21.377us 478.68 Mb 0 b 626
aten::_to_copy 0.01% 1.300ms 0.15% 12.973ms 23.376us 478.68 Mb 1.25 Kb 555
model_inference 0.31% 27.435ms 100.00% 8.775s 8.775s 159.56 Mb -467.16 Mb 1
aten::linear 0.01% 1.257ms 1.15% 101.083ms 523.748us 54.14 Mb 0 b 193
aten::addmm 33.55% 2.944s 34.48% 3.026s 1.995ms 54.14 Mb 54.14 Mb 1517
Torch-Compiled Region 11.15% 978.694ms 243.17% 21.338s 1.778s 49.11 Mb -5.99 Gb 12
aten::gelu 0.05% 4.537ms 0.05% 4.537ms 141.789us 24.06 Mb 24.06 Mb 32
aten::add 0.02% 2.115ms 0.02% 2.151ms 26.237us 19.46 Mb 19.46 Mb 82
aten::native_layer_norm 0.03% 2.407ms 0.03% 3.060ms 47.074us 12.24 Mb 0 b 65
aten::layer_norm 0.00% 175.379us 0.04% 3.235ms 49.772us 12.22 Mb -19.10 Kb 65
aten::div 0.00% 208.021us 0.00% 255.811us 23.256us 7.10 Mb 7.10 Mb 11
aten::cat 0.00% 164.204us 0.00% 188.948us 62.983us 7.05 Mb 7.05 Mb 3
----------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------
Self CPU time total: 8.775s
Loading

0 comments on commit b59670b

Please sign in to comment.