[PATCH] D49499: [X86] Prefer unpckhpd over movhlps in isel for fake unary cases

[Craig Topper via Phabricator via llvm-commits]

craig.topper added a comment.

@RKSimon what are your thoughts on this patch? Is this ok as is or should we prefer movhlps by default for code size? On Intel CPUs I think they are the same perf. I checked Haswell microcode and they use the same uop with the operands swapped.


https://reviews.llvm.org/D49499