[all-commits] [llvm/llvm-project] 026fac: [mlir][linalg] Vectorization for conv_1d_ncw_fcw

[Stanley Winata via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 026fac2a14cdf0b904ec83044d1f271e1ba2c5f9
      https://github.com/llvm/llvm-project/commit/026fac2a14cdf0b904ec83044d1f271e1ba2c5f9
  Author: Stanley Winata <stanley at nod-labs.com>
  Date:   2022-09-14 (Wed, 14 Sep 2022)

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

  Log Message:
  -----------
  [mlir][linalg] Vectorization for conv_1d_ncw_fcw

Most computer vision torch models uses nchw/ncw convolution. In a previous patch we added decomposition conv2dNchw to conv1dNcw. To enhance the performance on torch models we add this vectorization pattern for conv1dNcw which would consquently also improve the performance on conv2dNchw.

On IREE + Intel Xeon 8360 + Resnet50, we were able to get ~7x speed up ~880ms to 126ms.

Reviewed By: nicolasvasilache, hanchung

Differential Revision: https://reviews.llvm.org/D133675