composable_kernel

mirror of https://github.com/ROCm/composable_kernel.git synced 2026-06-28 18:56:59 +00:00

Files

Wojciech Laskowski c2601f38b7 [rocm-libraries] ROCm/rocm-libraries#6569 (commit 393049e)

Adding amdgcn_mma specializations for sparse MFMA builtins
 (#6569)

## Motivation

This PR is part of the [WMMA/MFMA] unification work. It's the fourth of
the series of PRs (after
https://github.com/ROCm/rocm-libraries/pull/5801,
https://github.com/ROCm/rocm-libraries/pull/6014 and
https://github.com/ROCm/rocm-libraries/pull/6567) that add all the
necessary MMA builtins as amdgcn_mma structs. This PR focuses on sparse
MFMA intrinsics.

## Technical Details

This change adds new specializations for MFMA sparse builtins. In total,
we add 27 MFMA builtins.

## Test Plan

All the new wrappers were added to the test suite in
`test_amdgcn_mma_layout.inc`.

## Test Result

Test pass locally, waiting for the CI.

## Submission Checklist

- [x] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

2026-06-12 12:48:29 +00:00

mma

[rocm-libraries] ROCm/rocm-libraries#6569 (commit 393049e)

2026-06-12 12:48:29 +00:00

amd_buffer_addressing_builtins.hpp

[rocm-libraries] ROCm/rocm-libraries#6768 (commit 43ca43f)

2026-06-05 12:27:41 +00:00

amd_buffer_addressing.hpp

[rocm-libraries] ROCm/rocm-libraries#6768 (commit 43ca43f)

2026-06-05 12:27:41 +00:00

amd_buffer_coherence.hpp

[rocm-libraries] ROCm/rocm-libraries#7836 (commit cdd9958)