Skip to content

Latest commit

 

History

History
54 lines (44 loc) · 3.72 KB

benchmark.md

File metadata and controls

54 lines (44 loc) · 3.72 KB
Raw

Setup

Tested 3 setups, pytorch 2.4

  1. AMD rx6600XT, OpenCL drivers vs official ROCM pytorch (6.1)
  2. NVidia rx960, OpenCL drivers vs official CUDA 12.2
  3. Inter Arc A380, OpenCL NEO driver vs XPU - intel extension for pytorch (2.1 since it what was released)

Input is standard Image net batchx3x224x224, time in milliseconds, lower is better.

Training

AMD Nvidia Intel
rx6600xt batch OpenCL ROCM % Perf gtx960 batch OpenCL CUDA % Perf A380 batch OpenCL XPU % Perf
alexnet 64 75.239 57.957 77 alexnet 64 257.09 130.561 51 alexnet 64 482.139 133.512 28
resnet18 64 238.927 147.099 62 resnet18 64 695.096 419.69 60 resnet18 64 1044.985 397.738 38
resnet50 32 358.872 266.155 74 resnet50 16 591.143 375.644 64 resnet50 16 640.916 329.849 51
convnext_small 16 608.297 337.736 56 convnext_small 8 1001.294 1120.676 112 convnext_small 8 841.302 259.292 31
vgg16 16 343.962 206.243 60 vgg16 8 520.75 363.288 70 vgg16 8 780.692 479.314 61
densenet161 16 494.175 297.001 60 densenet161 8 698.842 464.051 66 densenet161 8 834.207 423.883 51
mobilenet_v2 32 206.255 157.743 76 mobilenet_v2 16 335.279 173.748 52 mobilenet_v2 16 405.541 153.694 38
mobilenet_v3_small 64 130.571 92.83 71 mobilenet_v3_small 32 196.173 102.561 52 mobilenet_v3_small 32 275.302 92.086 33
mobilenet_v3_large 64 330.269 287.3 87 mobilenet_v3_large 32 497.168 264.072 53 mobilenet_v3_large 32 642.568 226.292 35
resnext50_32x4d 32 490.971 336.183 68 resnext50_32x4d 16 807.178 539.026 67 resnext50_32x4d 16 1068.918 396.39 37
wide_resnet50_2 32 643.083 468.04 73 wide_resnet50_2 16 1023.105 677.723 66 wide_resnet50_2 16 1373.346 634.213 46
mnasnet1_0 32 167.934 160.254 95 mnasnet1_0 16 302.854 167.911 55 mnasnet1_0 16 383.069 126.56 33
efficientnet_b0 32 313.972 205.674 66 efficientnet_b0 16 515.058 241.311 47 efficientnet_b0 16 531.724 203.157 38
regnet_y_400mf 64 246.069 171.841 70 regnet_y_400mf 32 361.507 353.584 98 regnet_y_400mf 32 635.279 224.228 35
Average 71 Average 65 Average 40

Inference

AMD Nvidia Intel
rx6600xt batch OpenCL ROCM % Perf gtx960 batch OpenCL CUDA % Perf A380 batch OpenCL XPU % Perf
alexnet 64 24.543 24.642 100 alexnet 32 45.007 30.271 67 alexnet 32 55.5 25.835 47
resnet18 64 59.428 41.569 70 resnet18 32 94.044 66.61 71 resnet18 32 113.002 55.647 49
resnet50 64 196.75 165.706 84 resnet50 32 316.899 215.245 68 resnet50 32 271.778 145.842 54
convnext_small 64 632.215 478.088 76 convnext_small 32 881.586 751.286 85 convnext_small 32 670.291 294.405 44
vgg16 64 310.767 205.745 66 vgg16 32 490.68 351.488 72 vgg16 32 801.684 333.954 42
densenet161 64 415.707 410.906 99 densenet161 32 589.712 510.883 87 densenet161 32 685.154 315.407 46
mobilenet_v2 64 93.699 77.774 83 mobilenet_v2 32 162.4 87.376 54 mobilenet_v2 32 100.363 51.589 51
mobilenet_v3_small 64 25.653 22.253 87 mobilenet_v3_small 32 50.097 28.739 57 mobilenet_v3_small 32 36.92 26.508 72
mobilenet_v3_large 64 70.409 63.28 90 mobilenet_v3_large 32 122.416 69.432 57 mobilenet_v3_large 32 84.413 52.328 62
resnext50_32x4d 64 274.967 245.411 89 resnext50_32x4d 32 440.411 284.571 65 resnext50_32x4d 32 359.037 169.194 47
wide_resnet50_2 64 404.214 321.398 80 wide_resnet50_2 32 589.164 376.938 64 wide_resnet50_2 32 682.184 321.014 47
mnasnet1_0 64 75.027 74.211 99 mnasnet1_0 32 133.324 83.407 63 mnasnet1_0 32 91.441 51.785 57
efficientnet_b0 64 114.735 104.417 91 efficientnet_b0 32 203.531 111.822 55 efficientnet_b0 32 129.755 88.131 68
regnet_y_400mf 64 57.408 43.313 75 regnet_y_400mf 32 96.079 99.022 103 regnet_y_400mf 32 87.756 56.503 64
Average 85 Average 69 Average 54