[clang] [X86] Add MMX/SSE/AVX PHADD/SUB & HADDPS/D intrinsics to be used in constexpr (PR #156822)

[via cfe-commits]

================
@@ -3453,6 +3592,30 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
   case Builtin::BI__builtin_elementwise_min:
     return interp__builtin_elementwise_maxmin(S, OpPC, Call, BuiltinID);
 
+  case clang::X86::BI__builtin_ia32_phaddw128:
+  case clang::X86::BI__builtin_ia32_phaddw256:
+  case clang::X86::BI__builtin_ia32_phaddd128:
+  case clang::X86::BI__builtin_ia32_phaddd256:
+  case clang::X86::BI__builtin_ia32_phaddsw128:
+  case clang::X86::BI__builtin_ia32_phaddsw256:
+  case clang::X86::BI__builtin_ia32_phsubw128:
+  case clang::X86::BI__builtin_ia32_phsubw256:
+  case clang::X86::BI__builtin_ia32_phsubd128:
+  case clang::X86::BI__builtin_ia32_phsubd256:
+  case clang::X86::BI__builtin_ia32_phsubsw128:
+  case clang::X86::BI__builtin_ia32_phsubsw256:
+    return interp_builtin_ia32ph_add_sub(S, OpPC, Frame, Call, BuiltinID);
+
+  case clang::X86::BI__builtin_ia32_haddpd:
+  case clang::X86::BI__builtin_ia32_haddpd256:
+  case clang::X86::BI__builtin_ia32_haddps:
+  case clang::X86::BI__builtin_ia32_haddps256:
+  case clang::X86::BI__builtin_ia32_hsubpd:
+  case clang::X86::BI__builtin_ia32_hsubpd256:
+  case clang::X86::BI__builtin_ia32_hsubps:
+  case clang::X86::BI__builtin_ia32_hsubps256:
+    return interp_builtin_floatph_add_sub(S, OpPC, Frame, Call, BuiltinID);
----------------
whytolearn wrote:

flaot operation involve rounding off mode [getRoundingMode], i think maybe use another function to  deal it is more better. 

https://github.com/llvm/llvm-project/pull/156822