-
Notifications
You must be signed in to change notification settings - Fork 3.2k
[WebGPU] Unify core implementations of GEMM and MatMul #24586
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
[WebGPU] Unify core implementations of GEMM and MatMul #24586
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
testing it with a bunch of models that use Gemm |
/azp run Linux QNN CI Pipeline,Win_TRT_Minimal_CUDA_Test_CI,Windows ARM64 QNN CI Pipeline,Windows GPU Doc Gen CI Pipeline,Windows x64 QNN CI Pipeline |
Azure Pipelines successfully started running 5 pipeline(s). |
The following test cases are failing on QNN EP:
|
Some CI jobs didn’t run. Trying to trigger them by close and re-open the PR. |
/azp run Linux QNN CI Pipeline,Win_TRT_Minimal_CUDA_Test_CI,Windows ARM64 QNN CI Pipeline,Windows GPU Doc Gen CI Pipeline,Windows x64 QNN CI Pipeline |
Azure Pipelines successfully started running 5 pipeline(s). |
/azp run Test Linux CUDA x64 Release,Test Linux TensorRT x64 Release,web_Debug / build_onnxruntime_web,web_Release / build_onnxruntime_web |
No pipelines are associated with this pull request. |
/azp run web_Release / build_onnxruntime_web |
No pipelines are associated with this pull request. |
I fear you need to merge with main - some definitions for the CI have changed :( |
It seems that Lint pipeline is broken with the latest main. |
There are some known issues and it's being fixed. please wait for the fixes merged into main. |
/azp run web_Debug / build_onnxruntime_web |
No pipelines are associated with this pull request. |
Merged latest main branch, which contains the CI pipeline fixes. |
/azp run Linux QNN CI Pipeline,Win_TRT_Minimal_CUDA_Test_CI,Windows ARM64 QNN CI Pipeline,Windows GPU Doc Gen CI Pipeline,Windows x64 QNN CI Pipeline |
Azure Pipelines successfully started running 5 pipeline(s). |
Description
This PR extract core implementations into gemm_utils.cc which is used to generate shader both GEMM and Matmul ops. The core implemenations included scalar and vec4 versions of GEMM and Matmul.
Motivation and Context
There are many common codes for GEMM and Matmul, so we want to extra common code to unify their implementations.
