[PATCH] D37947: [X86] Fix two more places to prefer VPERMQ/PD over VPERM2X128 when AVX2 is enabled
Simon Pilgrim via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Sun Sep 17 03:48:33 PDT 2017
RKSimon added inline comments.
================
Comment at: lib/Target/X86/X86ISelLowering.cpp:12150
+ Flipped, DAG.getConstant(PERMMask, DL, MVT::i8));
+ }
Flipped = DAG.getBitcast(VT, Flipped);
----------------
What happens if we just replace this with a generic DAG.getVectorShuffle? I'm thinking about when we finally get around to supporting 512-bit in lowerVectorShuffleAsLanePermuteAndBlend.
https://reviews.llvm.org/D37947
More information about the llvm-commits
mailing list