[all-commits] [llvm/llvm-project] 9b4752: Reorder linalg.conv indexing_maps loop order

[Ahmed S. Taei via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: 9b47525824df9be5ae23c39e7ce0d220d12f85e5
      https://github.com/llvm/llvm-project/commit/9b47525824df9be5ae23c39e7ce0d220d12f85e5
  Author: Ahmed S. Taei <ataei at google.com>
  Date:   2020-09-22 (Tue, 22 Sep 2020)

  Changed paths:
    M mlir/include/mlir/Dialect/Linalg/IR/LinalgStructuredOps.td
    M mlir/test/Dialect/Linalg/affine.mlir
    M mlir/test/Dialect/Linalg/loops.mlir
    M mlir/test/Dialect/Linalg/tile_conv.mlir

  Log Message:
  -----------
  Reorder linalg.conv indexing_maps loop order

Change the indexing map to iterate over the (b, x0, x1, z0, z1, q, k) instead of (b, x0, x1, k, q, z0, z1) to evaluate the convolution expression:
Y[b, x0, x1, k] = sum(W[z0, z1, q, k] * X[b, x0 + z0, x1 + z1, q], z0, z1, q)

This allows llvm auto vectorize to work and has better locality resulting significant performance improvments

Differential Revision: https://reviews.llvm.org/D87781