__fp_classNN
In this section:
Syntax
unsigned int __fp_class32(float);
unsigned int __fp_class64(double);
Description
Classifies a floating-point value, using the instruction fclass.s (32-bit FPU) or fclass.d (64-bit FPU). A compatible FPU must be available.
A bit mask with exactly one bit set is returned. The possible values are:
| Negative infinity |
| Negative normal number |
| Negative subnormal number |
| Negative zero |
| Positive infinity |
| Positive normal number |
| Positive subnormal number |
| Positive zero |
| Signaling NaN |
| Quiet NaN |
Example
if ( __fp_class32(value)
& ( _FP_NEGATIVE_SUBNORMAL
| _FP_POSITIVE_SUBNORMAL))
{
/* value is subnormal */
}Note
To use intrinsic functions in an application, you must include the header file(s) where they are declared, see Summary of intrinsic functions.