[PATCH] D72575: [x86] try harder to form 256-bit unpck*

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added inline comments.


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:16190
   if (V2.isUndef()) {
+    if (SDValue V = lowerShuffleWithUNPCK256(DL, MVT::v8i32, Mask, V1, V2, DAG))
+      return V;
----------------
Add a comment and move the VPERMD comment.


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:16251
+  if (SDValue V = lowerShuffleWithUNPCK256(DL, MVT::v16i16, Mask, V1, V2, DAG))
+    return V;
+
----------------
Add a comment.

This generates a shuffle sequence, its probably better to place this after all the lowerShuffle* calls that create single (non-variable) instructions.


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:16354
+  if (SDValue V = lowerShuffleWithUNPCK256(DL, MVT::v32i8, Mask, V1, V2, DAG))
+    return V;
+
----------------
Add a comment.

This generates a shuffle sequence, its probably better to place this after all the lowerShuffle* calls that create single (non-variable) instructions.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D72575/new/

https://reviews.llvm.org/D72575