[llvm] [AMDGPU] Add half vector support for table-driven libcall optimzation (PR #178638)

[Steffen Larsen via llvm-commits]

================
@@ -743,6 +743,38 @@ bool AMDGPULibCalls::fold(CallInst *CI) {
   return false;
 }
 
+static Constant *
+_Z4coshdgetConstantFloatVectorForArgType(LLVMContext &Ctx,
----------------
steffenlarsen wrote:

Hah! I did a last-minute renaming of the function, but must have had something in the "replace" field already. It has been fixed.

https://github.com/llvm/llvm-project/pull/178638