[X86][AVX] Eliminate redundant movzbl instruction.
authorLuo, Yuanke <yuanke.luo@intel.com>
Wed, 19 Oct 2022 00:32:13 +0000 (08:32 +0800)
committerLuo, Yuanke <yuanke.luo@intel.com>
Thu, 20 Oct 2022 23:17:51 +0000 (07:17 +0800)
commit889911ec268a342ec5da0e343eda85233fb62f5a
tree00d46f951b2908acac383769fc434435edda4385
parente874bbc2925cb1fc96f060ab681098219b7b2b28
[X86][AVX] Eliminate redundant movzbl instruction.

The movzbl instruction can be combined to vpinsrb or vmovd, when it is
actual lowered from anyext.

Differential Revision: https://reviews.llvm.org/D130953
llvm/lib/Target/X86/X86InstrAVX512.td
llvm/lib/Target/X86/X86InstrSSE.td
llvm/test/CodeGen/X86/avx-intrinsics-fast-isel.ll
llvm/test/CodeGen/X86/load-scalar-as-vector.ll
llvm/test/CodeGen/X86/pr15267.ll
llvm/test/CodeGen/X86/setcc-lowering.ll
llvm/test/CodeGen/X86/sse2-intrinsics-fast-isel.ll
llvm/test/CodeGen/X86/sse41-intrinsics-fast-isel.ll