The _ _SUBS_C intrinsic has the following declaration:
float __SUBS_C(float, float);
This intrinsic function returns the result of subtracting the second argument from the first argument, with IEEE chopped rounding used. Both arguments and the result are in IEEE single-precision format.