[PATCH] D12270: AVX512 : vpalignr implementation

[Elena Demikhovsky via llvm-commits]

delena added inline comments.

================
Comment at: lib/Target/X86/X86InstrSSE.td:5802
@@ -5801,3 +5801,3 @@
 
-let Predicates = [HasAVX2] in {
+let Predicates = [UseAVX2] in {
 def : Pat<(v8i32 (X86PAlignr VR256:$src1, VR256:$src2, (i8 imm:$imm))),
----------------
The same instruction is supported in AVX-512 VL/BW. I think you should specify this.
Please add IR tests for AVX-512 to see that the shuffles still work.


Repository:
  rL LLVM

http://reviews.llvm.org/D12270