[all-commits] [llvm/llvm-project] 664666: [mlir][sparse] Improve ConcatenateOp rewriting for...

[Bixia Zheng via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 66466641545a0b392703c0a81b50b9f31c01d54e
      https://github.com/llvm/llvm-project/commit/66466641545a0b392703c0a81b50b9f31c01d54e
  Author: bixia1 <bixia at google.com>
  Date:   2023-01-13 (Fri, 13 Jan 2023)

  Changed paths:
    M mlir/lib/Dialect/SparseTensor/Transforms/CodegenUtils.cpp
    M mlir/lib/Dialect/SparseTensor/Transforms/CodegenUtils.h
    M mlir/lib/Dialect/SparseTensor/Transforms/SparseTensorRewriting.cpp
    M mlir/test/Dialect/SparseTensor/sparse_concat_codegen.mlir

  Log Message:
  -----------
  [mlir][sparse] Improve ConcatenateOp rewriting for annotated all dense result.

Previously, we rely on InsertOp to add values to the result, in the same way we
add values to a sparse tensor with compressed dimensions. We now direct store
values to the values buffer.

Reviewed By: Peiming

Differential Revision: https://reviews.llvm.org/D141517