Skip to content

cpu: x64: matmul: enable brg for fp32 on avx2 for all non-gemv cases#4943

Open
azhai219 wants to merge 1 commit intouxlfoundation:rls-v3.10from
azhai219:az/rls-v3.10
Open

cpu: x64: matmul: enable brg for fp32 on avx2 for all non-gemv cases#4943
azhai219 wants to merge 1 commit intouxlfoundation:rls-v3.10from
azhai219:az/rls-v3.10

Conversation

@azhai219
Copy link
Copy Markdown

@azhai219 azhai219 commented Apr 2, 2026

Description

Please include a summary of the change. Please also include relevant motivation and context. See contribution guidelines for more details. If the change fixes an issue not documented in the project's Github issue tracker, please document all steps necessary to reproduce it.

Fixes # (github issue)

Checklist

General

  • Do all unit and benchdnn tests (make test and make test_benchdnn_*) pass locally for each commit?
  • Have you formatted the code using clang-format?

Performance improvements

  • Have you submitted performance data that demonstrates performance improvements?

New features

  • Have you published an RFC for the new feature?
  • Was the RFC approved?
  • Have you added relevant tests?

Bug fixes

  • Have you included information on how to reproduce the issue (either in a github issue or in this PR)?
  • Have you added relevant regression tests?

RFC PR

  • Does RFC document follow the template?
  • Have you added a link to the rendered document?

@azhai219 azhai219 requested a review from a team as a code owner April 2, 2026 05:58
@github-actions github-actions bot added platform:cpu-x64 Intel64/AMD64 processors. Codeowner: @oneapi-src/onednn-cpu-x64 backport labels Apr 2, 2026
// for the N=1 case, which is the only scenario guaranteed to
// perform on par or better than the GEMM implementation.
// For all other cases, we prefer to fall back to GEMM implementation,
if (!bgmmc.is_gemv && bm_conf_utils.is_f32() && bgmmc.isa == avx2
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  1. brgemm matmul: avx2: f32: closing performance gap for non-gemv cases #4097 - closes performance gap between gemm and brgemm matmuls for non-gemv cases
  2. cpu: x64: matmul fix wei_k_blk query #4134 - fix

So enabling brgemm matmul in rls-v3.10 by default without the optimizations will introduce regressions.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

backport platform:cpu-x64 Intel64/AMD64 processors. Codeowner: @oneapi-src/onednn-cpu-x64

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants