[PATCH] D73197: [x86] match concat of 128-bit high half vectors before lowering to VPERM2X128

[Sanjay Patel via Phabricator via llvm-commits]

spatel added a comment.

In D73197#1834183 <https://reviews.llvm.org/D73197#1834183>, @RKSimon wrote:

> This looks OK, but I'd be curious if we'd hit more cases by adding a vperm2f128(insert_subvector,insert_subvector) combine in combineTargetShuffle instead.


I drafted a combineTargetShuffle() alternative, and it doesn't cause any difference on existing regression tests. I'll post it here for comparison.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D73197/new/

https://reviews.llvm.org/D73197