[clang] [llvm] [clang][NVPTX] Add remaining float to fp16 conversions (PR #167641)

[Srinivasa Ravi via llvm-commits]

================
@@ -596,6 +596,15 @@ let hasSideEffects = false in {
   defm CVT_bf16 : CVT_FROM_ALL<"bf16", B16, [hasPTX<78>, hasSM<90>]>;
   defm CVT_f32 : CVT_FROM_ALL<"f32", B32>;
   defm CVT_f64 : CVT_FROM_ALL<"f64", B64>;
+  
+  multiclass CVT_FROM_FLOAT_SATFINITE<string ToName, RegisterClass RC> {
+    def _f32_sf :
----------------
Wolfram70 wrote:

Yeah, that was my thought as well so I'm thinking to keep this as a multiclass for now.

https://github.com/llvm/llvm-project/pull/167641