[all-commits] [llvm/llvm-project] 5ec628: [MLIR][AArch64] Refactor lowering of vector.contra...

Momchil Velikov via All-commits all-commits at lists.llvm.org
Fri Jul 11 03:35:49 PDT 2025


  Branch: refs/heads/users/momchil-velikov/vector-contract-bfmmla-neon
  Home:   https://github.com/llvm/llvm-project
  Commit: 5ec628726772f32ad2cb4ccfbbf0a43532562fca
      https://github.com/llvm/llvm-project/commit/5ec628726772f32ad2cb4ccfbbf0a43532562fca
  Author: Momchil Velikov <momchil.velikov at arm.com>
  Date:   2025-07-11 (Fri, 11 Jul 2025)

  Changed paths:
    M mlir/lib/Dialect/ArmNeon/Transforms/LowerContractionToNeonI8MMPattern.cpp

  Log Message:
  -----------
  [MLIR][AArch64] Refactor lowering of vector.contract to Neon I8MM

This patch refactors the pattern in `Transforms/LowerContractionToNeonI8MMPattern.cpp`
using similar approach as in https://github.com/llvm/llvm-project/pull/147052
to prepare for adding BF16 support.


  Commit: 442e29ac0b3ad54aa521ff02029889fa899a9c75
      https://github.com/llvm/llvm-project/commit/442e29ac0b3ad54aa521ff02029889fa899a9c75
  Author: Momchil Velikov <momchil.velikov at arm.com>
  Date:   2025-07-11 (Fri, 11 Jul 2025)

  Changed paths:
    M mlir/include/mlir/Conversion/Passes.td
    M mlir/include/mlir/Dialect/ArmNeon/TransformOps/ArmNeonVectorTransformOps.td
    M mlir/include/mlir/Dialect/ArmNeon/Transforms.h
    M mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
    M mlir/lib/Dialect/ArmNeon/TransformOps/ArmNeonVectorTransformOps.cpp
    M mlir/lib/Dialect/ArmNeon/Transforms/CMakeLists.txt
    A mlir/lib/Dialect/ArmNeon/Transforms/LowerContractToNeonPatterns.cpp
    R mlir/lib/Dialect/ArmNeon/Transforms/LowerContractionToNeonI8MMPattern.cpp
    M mlir/lib/Dialect/ArmSVE/Transforms/LowerContractionToSVEI8MMPattern.cpp
    A mlir/test/Dialect/ArmNeon/vector-bfmmla.mlir
    A mlir/test/Integration/Dialect/Vector/CPU/ArmNeon/vector-contract-bfmmla.mlir
    M mlir/test/Integration/Dialect/Vector/CPU/ArmNeon/vector-contract-i8mm.mlir

  Log Message:
  -----------
  [MLIR][AArch64] Lower vector.contract to Neon FEAT_BF16 operations


Compare: https://github.com/llvm/llvm-project/compare/5ec628726772%5E...442e29ac0b3a

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