[all-commits] [llvm/llvm-project] 6c59c5: [mlir][Vector] Re-define masking semantics in vect...

Diego Caballero via All-commits all-commits at lists.llvm.org
Thu Nov 17 17:07:00 PST 2022


  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 6c59c5cd08c879c9d1cfa653711613244a7c39bf
      https://github.com/llvm/llvm-project/commit/6c59c5cd08c879c9d1cfa653711613244a7c39bf
  Author: Diego Caballero <diegocaballero at google.com>
  Date:   2022-11-18 (Fri, 18 Nov 2022)

  Changed paths:
    M mlir/include/mlir/Dialect/Vector/IR/VectorOps.td
    M mlir/include/mlir/Interfaces/VectorInterfaces.h
    M mlir/include/mlir/Interfaces/VectorInterfaces.td
    M mlir/lib/Dialect/Vector/IR/VectorOps.cpp
    M mlir/lib/Dialect/Vector/Transforms/VectorTransferPermutationMapRewritePatterns.cpp
    M mlir/lib/Interfaces/VectorInterfaces.cpp
    M mlir/test/Conversion/VectorToSCF/vector-to-scf-mask-and-permutation-map.mlir
    M mlir/test/Dialect/Vector/ops.mlir
    M mlir/test/Dialect/Vector/vector-transfer-to-vector-load-store.mlir

  Log Message:
  -----------
  [mlir][Vector] Re-define masking semantics in vector.transfer ops

Masking hasn't been widely used in vector transfer ops and the semantics
of the mask operand were a bit loose. This patch states that the mask
operand in a vector transfer op is applied to the read/write part of the
operation and, therefore, its shape should match the shape of the
elements read/written from/into the memref/tensor regardless of any
permutation/broadcasting also applied by the transfer operation.

Reviewers: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D138079




More information about the All-commits mailing list