[all-commits] [llvm/llvm-project] 6bb0ab: [MLIR] Propagate unpack through element-wise ops

[lorenzo chelini via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 6bb0ab0de039ddbcc70eafc549b50da7867fb617
      https://github.com/llvm/llvm-project/commit/6bb0ab0de039ddbcc70eafc549b50da7867fb617
  Author: Lorenzo Chelini <l.chelini at icloud.com>
  Date:   2023-02-01 (Wed, 01 Feb 2023)

  Changed paths:
    M mlir/lib/Dialect/Linalg/Transforms/DataLayoutPropagation.cpp
    M mlir/test/Dialect/Linalg/data-layout-propagation.mlir

  Log Message:
  -----------
  [MLIR] Propagate unpack through element-wise ops

Introduce `pushDownUnPackOpThroughElemGenericOp` to propagate producer
unpack operation through an element-wise linalg.generic operation. This
pattern complements `BubbleUpPackOpThroughElemGenericOp`. The general
idea is to bubble up tensor.pack as much as possible while pushing down
tensor.unpack as much as possible, and canonicalize away symmetrical
tensor.pack and tensor.unpack operations.

Currently, `pushDownUnPackOpThroughElemGenericOp` expects a single
tensor.unpack operation as the producer of one of the linalg.generic's
operands.

Reviewed By: hanchung

Differential Revision: https://reviews.llvm.org/D142523