[all-commits] [llvm/llvm-project] 6292ea: [mlir][AMDGPU] Remove an old bf16 workaround (#108...

Krzysztof Drewniak via All-commits all-commits at lists.llvm.org
Thu Sep 12 15:46:00 PDT 2024


  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 6292ea6879217468cd9187d4f4dd3ee7c713431c
      https://github.com/llvm/llvm-project/commit/6292ea6879217468cd9187d4f4dd3ee7c713431c
  Author: Krzysztof Drewniak <Krzysztof.Drewniak at amd.com>
  Date:   2024-09-12 (Thu, 12 Sep 2024)

  Changed paths:
    M mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp
    M mlir/test/Conversion/AMDGPUToROCDL/wmma.mlir
    M mlir/test/Conversion/GPUToROCDL/gpu-to-rocdl.mlir

  Log Message:
  -----------
  [mlir][AMDGPU] Remove an old bf16 workaround (#108409)

The AMDGPU backend now implements LLVM's `bfloat` type. Therefore, we no
longer need to type convert MLIR's `bf16` to `i16` during lowerings to
ROCDL.

As a result of this change, we discovered that, whel the code for MFMA
and WMMA intrinsics was mainly prepared for this change, we were failing
to bitcast the bf16 results of WMMA operations out from the i16 they're
natively represented as. This commit also fixes that issue.

---------

Co-authored-by: Jakub Kuderski <kubakuderski at gmail.com>



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications


More information about the All-commits mailing list