projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: a993bb0) | patch
[mlir][sparse] support new kind of scalar in sparse linalg generic op
authorAart Bik <ajcbik@google.com>
Tue, 15 Jun 2021 22:56:32 +0000 (15:56 -0700)
committerAart Bik <ajcbik@google.com>
Wed, 16 Jun 2021 18:00:49 +0000 (11:00 -0700)
commit619bfe8bd23f76b22f0a53fedafbfc8c97a15f12
tree7792cd5e28d5744cb99ab1c0aa0557a489d697cdtree | snapshot
parenta993bb08b8348c947a733377655af3de610cf28ecommit | diff
[mlir][sparse] support new kind of scalar in sparse linalg generic op

We have several ways of introducing a scalar invariant value into
linalg generic ops (should we limit this somewhat?). This revision
makes sure we handle all of them correctly in the sparse compiler.

Reviewed By: gysit

Differential Revision: https://reviews.llvm.org/D104335
mlir/lib/Dialect/SparseTensor/Transforms/Sparsification.cpp diff | blob | history
mlir/test/Dialect/SparseTensor/sparse_scalars.mlir [new file with mode: 0644] blob
Domain: System / Toolchain;