Prioritize lowering V{4|16}F32 with blend.
authorNoah Goldstein <goldstein.w.n@gmail.com>
Thu, 16 Feb 2023 17:55:58 +0000 (11:55 -0600)
committerNoah Goldstein <goldstein.w.n@gmail.com>
Fri, 24 Feb 2023 21:22:08 +0000 (15:22 -0600)
commitaca34da46da41792614799a8b6a8b31a5a6e23d9
tree00a015dae4a58ae3ddaca8134d46deec72b8cb39
parentf1c8b72c13f12b8711f652291d434f921a5c9a66
Prioritize lowering V{4|16}F32 with blend.

Blend is often fastest available instruction so it should be higher
priority for v4f32 and an option for v16f32.

Reviewed By: RKSimon

Differential Revision: https://reviews.llvm.org/D143856
llvm/lib/Target/X86/X86ISelLowering.cpp
llvm/test/CodeGen/X86/avx512-insert-extract.ll
llvm/test/CodeGen/X86/merge-consecutive-loads-512.ll