[PATCH] D143789: [X86] Widen i16 shuffle masks if vector width < 512 even with BWI
Noah Goldstein via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Sun Feb 12 14:25:48 PST 2023
goldstein.w.n added a comment.
In D143789#4120191 <https://reviews.llvm.org/D143789#4120191>, @RKSimon wrote:
> lowerVECTOR_SHUFFLE should always widen a shuffle as much as possible - my guess is canCombineAsMaskOperation is blocking it on AVX512BW targets - maybe address it there instead?
Done in V2.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D143789/new/
https://reviews.llvm.org/D143789
More information about the llvm-commits
mailing list