[all-commits] [llvm/llvm-project] 414293: [mlir][Linalg] Move named op conversions out of ca...

[MaheshRavishankar via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 4142932a834f0dca9e9ae0c3754f097ffa3fc1ef
      https://github.com/llvm/llvm-project/commit/4142932a834f0dca9e9ae0c3754f097ffa3fc1ef
  Author: MaheshRavishankar <ravishankarm at google.com>
  Date:   2021-12-20 (Mon, 20 Dec 2021)

  Changed paths:
    M mlir/include/mlir/Dialect/Linalg/Passes.h
    M mlir/include/mlir/Dialect/Linalg/Passes.td
    M mlir/include/mlir/Dialect/Linalg/Transforms/Transforms.h
    M mlir/lib/Dialect/Linalg/IR/LinalgOps.cpp
    M mlir/lib/Dialect/Linalg/Transforms/CMakeLists.txt
    A mlir/lib/Dialect/Linalg/Transforms/NamedOpConversions.cpp
    M mlir/lib/Dialect/Linalg/Transforms/PassDetail.h
    M mlir/test/Dialect/Linalg/canonicalize.mlir
    A mlir/test/Dialect/Linalg/namedop_conversion.mlir

  Log Message:
  -----------
  [mlir][Linalg] Move named op conversions out of canonicalizations.

These conversions are better suited to be applied at whole tensor
level. Applying these as canonicalizations end up triggering such
canonicalizations at all levels of the stack which might be
undesirable. For example some of the resulting code patterns wont
bufferize in-place and need additional stack buffers. Best is to be
more deliberate in when these canonicalizations apply.

Differential Revision: https://reviews.llvm.org/D115912