projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 318ce4a) | patch
[mlir][linalg] Improve codegen when tiling PadTensor evenly
authorMatthias Springer <springerm@google.com>
Thu, 15 Jul 2021 02:27:52 +0000 (11:27 +0900)
committerMatthias Springer <springerm@google.com>
Thu, 15 Jul 2021 02:29:21 +0000 (11:29 +0900)
commita0e02018beb81946397f577f14df09e4b3b675da
treed474c4df203bac2025c70004269958b075596e04tree | snapshot
parent318ce4ad927d129a2bf96c2c872f4d107c45bdefcommit | diff
[mlir][linalg] Improve codegen when tiling PadTensor evenly

Produce simpler IR with more static type information and fewer affine expressions.

Differential Revision: https://reviews.llvm.org/D105530
mlir/lib/Dialect/Linalg/Transforms/Tiling.cpp diff | blob | history
mlir/test/Dialect/Linalg/tile-pad-tensor-op.mlir diff | blob | history
mlir/test/Dialect/Linalg/tile.mlir diff | blob | history
Domain: System / Toolchain;