projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 3c4d2a0) | patch
[mlir][sparse][vectorization] optimize reduction chains
authorAart Bik <ajcbik@google.com>
Tue, 22 Nov 2022 18:49:12 +0000 (10:49 -0800)
committerAart Bik <ajcbik@google.com>
Sat, 26 Nov 2022 20:40:51 +0000 (12:40 -0800)
commitcb82d375a8060bd3af83b64d7d2c94f4a59d4b97
tree0662f7ac24a6bea6c30bed2557fa9c208e6ba419tree | snapshot
parent3c4d2a03968ccf5889bacffe02d6fa2443b0260fcommit | diff
[mlir][sparse][vectorization] optimize reduction chains

A few more dots on the i's of the sparse vectorizer.
Also makes reduction matching less brittle.

Reviewed By: qcolombet

Differential Revision: https://reviews.llvm.org/D138513
mlir/lib/Dialect/SparseTensor/Transforms/SparseVectorization.cpp diff | blob | history
mlir/test/Dialect/SparseTensor/sparse_vector_chain.mlir [new file with mode: 0644] blob
Domain: System / Toolchain;