[PATCH] D31509: [InstCombine] Combine vector shuffles if the same operand can be reused

[Eli Friedman via Phabricator via llvm-commits]

efriedma added a comment.

x86 might have improved, but we still don't have a good story in general for handling arbitrary shuffles for large vector types with limited shuffle support.

I would be okay with poking a hole for shuffles which are equivalent to EXTRACT_SUBVECTOR, if that helps here.


https://reviews.llvm.org/D31509