[all-commits] [llvm/llvm-project] d271fc: [mlir][gpu] Split ops sinking from gpu-kernel-outl...

[Ivan Butygin via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: d271fc04d5b97b12e6b797c6067d3c96a8d7470e
      https://github.com/llvm/llvm-project/commit/d271fc04d5b97b12e6b797c6067d3c96a8d7470e
  Author: Ivan Butygin <ivan.butygin at gmail.com>
  Date:   2022-02-17 (Thu, 17 Feb 2022)

  Changed paths:
    M mlir/include/mlir/Dialect/GPU/Passes.h
    M mlir/include/mlir/Dialect/GPU/Passes.td
    M mlir/lib/Dialect/GPU/Transforms/KernelOutlining.cpp
    M mlir/lib/Dialect/GPU/Transforms/PassDetail.h
    M mlir/test/Dialect/GPU/outlining.mlir
    A mlir/test/Dialect/GPU/sink-ops.mlir

  Log Message:
  -----------
  [mlir][gpu] Split ops sinking from gpu-kernel-outlining pass into separate pass

Previously `gpu-kernel-outlining` pass was also doing index computation sinking into gpu.launch before actual outlining.
Split ops sinking from `gpu-kernel-outlining` pass into separate pass, so users can use theirs own sinking pass before outlining.
To achieve old behavior users will need to call both passes: `-gpu-launch-sink-index-computations -gpu-kernel-outlining`.

Differential Revision: https://reviews.llvm.org/D119932