These floating point rounding operations all use the roundss/roundsd instructions that are available in SSE 4.1.