[all-commits] [llvm/llvm-project] 641fe7: [mlir][Linalg] Fix and improve vectorization of de...

[Nicolas Vasilache via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 641fe70776c8fb0eba5c28a4aaa4a222e184a11d
      https://github.com/llvm/llvm-project/commit/641fe70776c8fb0eba5c28a4aaa4a222e184a11d
  Author: Nicolas Vasilache <nicolas.vasilache at gmail.com>
  Date:   2021-11-15 (Mon, 15 Nov 2021)

  Changed paths:
    M mlir/lib/Dialect/Linalg/Transforms/Vectorization.cpp
    M mlir/test/Dialect/Linalg/vectorize-convolution.mlir

  Log Message:
  -----------
  [mlir][Linalg] Fix and improve vectorization of depthwise convolutions.

When trying to connect the vectorization of depthwise convolutions to e2e execution
a number of problems surfaced.
Fix an off-by-one error on the size of the input vector (similary to what was previously done for regular conv).
Rewrite the lowering to vector.fma instead of vector.contract: the KW reduction dimension has already been unrolled and vector.contract requires a reduction dimension to be valid.

Differential Revision: https://reviews.llvm.org/D113884