[clang] [llvm] [HLSL] Implement dot2add intrinsic (PR #131237)

[Ashley Coleman via llvm-commits]

================
@@ -0,0 +1,8 @@
+; RUN: opt -S -dxil-op-lower -mtriple=dxil-pc-shadermodel6.3-compute %s | FileCheck %s
+
+define noundef float @dot2add_simple(<2 x half> noundef %a, <2 x half> noundef %b, float %c) {
+entry:
+; CHECK: call float @dx.op.dot2AddHalf(i32 162, float %c, half %0, half %1, half %2, half %3)
+  %ret = call float @llvm.dx.dot2add(<2 x half> %a, <2 x half> %b, float %c)
----------------
V-FEXrt wrote:

https://github.com/llvm/wg-hlsl/blob/main/proposals/0014-consistent-naming-for-dx-intrinsics.md is the official recommendation for stuff like this but it doesn't seem to comment on this one.

Does it make sense for there to ever be a "float" dot2add? If so, then I'd lean towards adding half to the name. But if this is the only spelling that makes sense then it's probably fine to drop it since the types are all included in the call

https://github.com/llvm/llvm-project/pull/131237