[PATCH] D47979: [X86] Lowering Mask Scalar add/sub/mul/div intrinsics to native IR (Clang part)

[Tomasz Krupa via Phabricator via cfe-commits]

tkrupa added inline comments.


================
Comment at: lib/CodeGen/CGBuiltin.cpp:9926
+    Value *Div = Builder.CreateFDiv(A, B);
+    llvm::VectorType *MaskTy = llvm::VectorType::get(Builder.getInt1Ty(),
+                             cast<IntegerType>(Mask->getType())->getBitWidth());
----------------
craig.topper wrote:
> Can we just emit the and+icmp that the other operations end up with?
We can't - if select condition is a CmpInst, CodeGenPrepare::optimizeSelectInst replaces it with a branch condition in case of expensive operations such as div. That's the reason I'm handling it in CGBuiltin in the first place.


Repository:
  rC Clang

https://reviews.llvm.org/D47979