[PATCH] D16691: [InstCombine] simplify masked load intrinsics with constant masks

[Sanjay Patel via llvm-commits]

spatel added inline comments.

================
Comment at: lib/Transforms/InstCombine/InstCombineCalls.cpp:680
@@ -679,3 +679,3 @@
 
 static Value *simplifyMinnumMaxnum(const IntrinsicInst &II) {
   Value *Arg0 = II.getArgOperand(0);
----------------
Thanks. I think the verifier has a hole - it doesn't confirm that the alignment operand is constant. But I've removed the assertions here because it does confirm the other things.

================
Comment at: lib/Transforms/InstCombine/InstCombineCalls.cpp:907
@@ +906,3 @@
+  // case Intrinsic::masked_store:
+  // case Intrinsic::masked_gather:
+  // case Intrinsic::masked_scatter:
----------------
Please correct me if I've misunderstood, but I think scatter and gather each have one degenerate folding opportunity that we can handle here:
1. A scatter with zero mask is a nop.
2. A gather with zero mask will return the passthru arg.


http://reviews.llvm.org/D16691