[PATCH] D29690: [AVX512] Fix EXTRACT_VECTOR_ELT for v2i1/v4i1/v32i1/v64i1 with variable index.

[Zvi Rackover via Phabricator via llvm-commits]

zvi added a comment.

Does it make sense to split this into 2 patches?

1. Replace X86ISD::VINSERT with BUILD_VECTOR
2. Implement missing functionality for variable-index extractelt?



================
Comment at: lib/Target/X86/X86ISelLowering.cpp:13658
 
+bool isVPERMVsupported(MVT VT, const X86Subtarget &Subtarget) {
+  if ((Subtarget.hasAVX2() && (VT == MVT::v8f32 || VT == MVT::v8i32)) ||
----------------
Make this a static function or consider making a member of X86Subtarget


Repository:
  rL LLVM

https://reviews.llvm.org/D29690