[all-commits] [llvm/llvm-project] 85aa5f: [mlir][SCF][GPU] Add DeviceMaskingAttrInterface su...

[Nicolas Vasilache via All-commits]

  Branch: refs/heads/users/nico/map-to-lanes-3
  Home:   https://github.com/llvm/llvm-project
  Commit: 85aa5f8c72801f5a75142a663d6e89e83e63decc
      https://github.com/llvm/llvm-project/commit/85aa5f8c72801f5a75142a663d6e89e83e63decc
  Author: Nicolas Vasilache <nico.vasilache at amd.com>
  Date:   2025-07-03 (Thu, 03 Jul 2025)

  Changed paths:
    M mlir/include/mlir/Dialect/GPU/IR/GPUDeviceMappingAttr.td
    M mlir/include/mlir/Dialect/GPU/TransformOps/Utils.h
    M mlir/include/mlir/Dialect/SCF/IR/DeviceMappingInterface.td
    M mlir/include/mlir/Dialect/SCF/IR/SCFOps.td
    M mlir/lib/Dialect/GPU/CMakeLists.txt
    M mlir/lib/Dialect/GPU/IR/GPUDialect.cpp
    M mlir/lib/Dialect/GPU/TransformOps/GPUTransformOps.cpp
    M mlir/lib/Dialect/GPU/TransformOps/Utils.cpp
    M mlir/lib/Dialect/SCF/IR/SCF.cpp
    M mlir/test/Dialect/GPU/transform-gpu-failing.mlir
    M mlir/test/Dialect/GPU/transform-gpu.mlir
    M mlir/test/Dialect/SCF/invalid.mlir

  Log Message:
  -----------
  [mlir][SCF][GPU] Add DeviceMaskingAttrInterface support to scf::ForallOp and use it to implement warp specialization.

This revision adds DeviceMaskingAttrInterface and extends
DeviceMappingArrayAttr to accept a union of DeviceMappingAttrInterface
and DeviceMaskingAttrInterface.

The first implementation is if the form of a GPUMappingMaskAttr, which
can be additionally passed to the scf.forall.mapping attribute to
specify a mask on compute resources that should be active.

Support is added to GPUTransformOps to take advantage of this
information and lower to block/warpgroup/warp/thread specialization when
mapped to linear ids.

Co-authored-by: Oleksandr "Alex" Zinenko <git at ozinenko.com>



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