Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add CK GEMM-Softmax-GEMM fusion #2250

Merged
merged 49 commits into from
Oct 13, 2023
Merged

Add CK GEMM-Softmax-GEMM fusion #2250

merged 49 commits into from
Oct 13, 2023

Conversation

turneram
Copy link
Contributor

No description provided.

@migraphx-bot
Copy link
Collaborator

migraphx-bot commented Sep 27, 2023

Test Batch Rate new
883732
Rate old
1a1c1b
Diff Compare
torchvision-resnet50 64 2,323.99 2,325.39 -0.06%
torchvision-resnet50_fp16 64 5,358.70 5,347.82 0.20%
torchvision-densenet121 32 1,847.03 1,849.21 -0.12%
torchvision-densenet121_fp16 32 3,410.48 3,406.11 0.13%
torchvision-inceptionv3 32 1,295.72 1,293.29 0.19%
torchvision-inceptionv3_fp16 32 2,536.60 2,537.00 -0.02%
cadene-inceptionv4 16 620.02 620.90 -0.14%
cadene-resnext64x4 16 588.80 588.26 0.09%
slim-mobilenet 64 7,219.83 7,213.68 0.09%
slim-nasnetalarge 64 236.41 236.58 -0.07%
slim-resnet50v2 64 2,559.45 2,557.33 0.08%
bert-mrpc-onnx 8 824.99 824.42 0.07%
bert-mrpc-tf 1 387.19 388.84 -0.43%
pytorch-examples-wlang-gru 1 299.70 298.85 0.29%
pytorch-examples-wlang-lstm 1 316.36 314.28 0.66%
torchvision-resnet50_1 1 543.92 549.39 -1.00%
torchvision-inceptionv3_1 1 305.91 302.45 1.14%
cadene-dpn92_1 1 353.11 353.66 -0.16%
cadene-resnext101_1 1 219.09 218.95 0.06%
slim-vgg16_1 1 224.68 224.19 0.22%
slim-mobilenet_1 1 1,495.14 1,510.80 -1.04%
slim-inceptionv4_1 1 216.37 216.72 -0.16%
onnx-taau-downsample 1 306.35 306.03 0.11%
dlrm-criteoterabyte 1 21.68 21.69 -0.03%
dlrm-criteoterabyte_fp16 1 40.75 40.72 0.06%
agentmodel 1 5,869.23 5,795.96 1.26%
unet_fp16 2 55.16 55.22 -0.10%
resnet50v1_fp16 1 751.35 761.81 -1.37%
bert_base_cased_fp16 64 970.93 971.55 -0.06%
bert_large_uncased_fp16 32 305.32 305.24 0.02%
bert_large_fp16 1 167.18 166.66 0.31%
distilgpt2_fp16 16 1,278.19 1,279.47 -0.10%

This build is OK for merge ✅

@migraphx-bot
Copy link
Collaborator

migraphx-bot commented Sep 27, 2023


    :white_check_mark:bert-mrpc-onnx: PASSED: MIGraphX meets tolerance

    :white_check_mark:bert-mrpc-tf: PASSED: MIGraphX meets tolerance

    :white_check_mark:pytorch-examples-wlang-gru: PASSED: MIGraphX meets tolerance

    :white_check_mark:pytorch-examples-wlang-lstm: PASSED: MIGraphX meets tolerance

    :white_check_mark:torchvision-resnet50_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:torchvision-inceptionv3_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:cadene-dpn92_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:cadene-resnext101_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:slim-vgg16_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:slim-mobilenet_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:slim-inceptionv4_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:dlrm-criteoterabyte: PASSED: MIGraphX meets tolerance

    :white_check_mark:agentmodel: PASSED: MIGraphX meets tolerance

    :white_check_mark:unet: PASSED: MIGraphX meets tolerance

    :white_check_mark:resnet50v1: PASSED: MIGraphX meets tolerance

🔴bert_base_cased_fp16: FAILED: MIGraphX is not within tolerance - check verbose output


🔴bert_large_uncased_fp16: FAILED: MIGraphX is not within tolerance - check verbose output


    :white_check_mark:bert_large: PASSED: MIGraphX meets tolerance

🔴distilgpt2_fp16: FAILED: MIGraphX is not within tolerance - check verbose output

src/include/migraphx/ck.hpp Outdated Show resolved Hide resolved
src/include/migraphx/ck.hpp Outdated Show resolved Hide resolved
src/targets/gpu/target.cpp Outdated Show resolved Hide resolved
@causten causten merged commit 05e3659 into develop Oct 13, 2023
15 checks passed
@causten causten deleted the ck-flash-attn branch October 13, 2023 17:43
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants