diff options
| author | Matthias Springer <springerm@google.com> | 2022-06-10 10:17:07 +0200 |
|---|---|---|
| committer | Matthias Springer <springerm@google.com> | 2022-06-10 10:26:07 +0200 |
| commit | 79f115911eebbd3c8f7758135e2533e041b0c2a3 (patch) | |
| tree | a16ca6ddec44da395f1b035746f0c7fc9e77b23b /llvm/lib/Transforms/Utils/LoopUtils.cpp | |
| parent | 914e30ca5c74643c4d90a4fd67845f9260de0b7d (diff) | |
| download | llvm-79f115911eebbd3c8f7758135e2533e041b0c2a3.zip llvm-79f115911eebbd3c8f7758135e2533e041b0c2a3.tar.gz llvm-79f115911eebbd3c8f7758135e2533e041b0c2a3.tar.bz2 | |
[mlir][bufferize] Avoid tensor copies when the data is not read
There are various shortcuts in `BufferizationState::getBuffer` that avoid a buffer copy when we just need an allocation (and no initialization). This change adds those shortcuts to the TensorCopyInsertion pass, so that `getBuffer` can be simplified in a subsequent change.
Differential Revision: https://reviews.llvm.org/D126821
Diffstat (limited to 'llvm/lib/Transforms/Utils/LoopUtils.cpp')
0 files changed, 0 insertions, 0 deletions