diff options
author | Peiming Liu <peiming@google.com> | 2022-10-28 20:45:06 +0000 |
---|---|---|
committer | Peiming Liu <peiming@google.com> | 2022-10-29 01:00:05 +0000 |
commit | e5cb0ee3836aacd8cc377a5a91659fc7512909c9 (patch) | |
tree | 3d3fb9727b3df194cd2f7040c1f05821d5813fae /clang/lib/Frontend/DependencyFile.cpp | |
parent | b9255099382d0121e567979e7d5b4ca1328b4a41 (diff) | |
download | llvm-e5cb0ee3836aacd8cc377a5a91659fc7512909c9.zip llvm-e5cb0ee3836aacd8cc377a5a91659fc7512909c9.tar.gz llvm-e5cb0ee3836aacd8cc377a5a91659fc7512909c9.tar.bz2 |
[mlir][sparse] run canonicalization pass after DenseOpBufferize.
As pointed out by Matthias: "DenseBufferizationPass should be run right after TensorCopyInsertionPass. (Running it after bufferizing the sparse IR is also OK.) The reason for this is that whether copies are needed for not depends on the structure of the program (SSA use-def chains). In particular, running the canonicalizer in-between is problematic because it could introduce new RaW conflicts"
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D136980
Diffstat (limited to 'clang/lib/Frontend/DependencyFile.cpp')
0 files changed, 0 insertions, 0 deletions