[PATCH] D11934: [InstCombine] Move SSE/AVX vector blend folding to instcombiner

[Andrea Di Biagio via llvm-commits]

andreadb accepted this revision.
andreadb added a comment.
This revision is now accepted and ready to land.

Hi Simon,

The patch LGTM.
Happy to see that we can now get rid of the target combine on INTRINSIC_WO_CHAIN.


================
Comment at: lib/Transforms/InstCombine/InstCombineCalls.cpp:968
@@ +967,3 @@
+
+    // fold (blend Op0, Op0, Mask) -> A
+    if (Op0 == Op1)
----------------
You mean:
  // fold (blend A, A, Mask) -> A


Repository:
  rL LLVM

http://reviews.llvm.org/D11934