[clang] [X86][Clang] Allow constexpr evaluation of F16C CVTPS2PH intrinsics (PR #162295)

[Hanyang Xu via cfe-commits]

================
@@ -3002,6 +3002,91 @@ static bool interp__builtin_vec_set(InterpState &S, CodePtr OpPC,
   return true;
 }
 
+static bool interp__builtin_ia32_vcvtps2ph(InterpState &S, CodePtr OpPC,
+                                           const CallExpr *Call) {
+  // Arguments are: vector of floats, rounding immediate
----------------
ericxu233 wrote:

Yes, since this is converting single-precision (float32) values to half-precision (float16) which is stored by __m128i.

https://github.com/llvm/llvm-project/pull/162295