[all-commits] [llvm/llvm-project] 2ef416: [mlir][sparse] Improve sort operation by generatin...

[Bixia Zheng via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 2ef416273f85ed355f4cbaa996fd8d9c229dbfab
      https://github.com/llvm/llvm-project/commit/2ef416273f85ed355f4cbaa996fd8d9c229dbfab
  Author: bixia1 <bixia at google.com>
  Date:   2023-03-14 (Tue, 14 Mar 2023)

  Changed paths:
    M mlir/lib/Dialect/SparseTensor/Transforms/SparseBufferRewriting.cpp
    M mlir/test/Dialect/SparseTensor/buffer_rewriting.mlir

  Log Message:
  -----------
  [mlir][sparse] Improve sort operation by generating inlined code to compare values.

Previously, we generate function calls to compare values for sorting. It turns
out that the compiler doesn't inline those function calls. We now directly
generate inlined code. Also, modify the code for comparing values to use less
number of branches.

This improves all sort implementation in general. For arabic-2005.mtx CSR, the
improvement is around 25%.

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D145442