projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 32c26e2) | patch
[mlir][vector] Disable folding for masked reductions
authorDiego Caballero <diegocaballero@google.com>
Thu, 19 Jan 2023 22:59:02 +0000 (22:59 +0000)
committerDiego Caballero <diegocaballero@google.com>
Thu, 19 Jan 2023 23:06:53 +0000 (23:06 +0000)
commit4626bd0b91102af234125f7f8ff0daaffb7a1fa4
tree7fcc4734bbdd49d6def36fa39ff8114ca5522527tree | snapshot
parent32c26e27b6fcd12703dcd00adf178330d0ad8449commit | diff
[mlir][vector] Disable folding for masked reductions

Reductions can't be folded into plain arith ops until we can mask
those arith ops.

Reviewed By: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D141645
mlir/lib/Dialect/Vector/IR/VectorOps.cpp diff | blob | history
mlir/test/Dialect/Vector/canonicalize.mlir diff | blob | history
Domain: System / Toolchain;