FP_NORMAL, FP_SUBNORMAL, FP_ZERO, FP_INFINITE, FP_NAN

From cppreference.com
< c‎ | numeric‎ | math
 
 
 
Common mathematical functions
Functions
Basic operations
(C99)
(C99)
(C99)
(C99)(C99)(C99)(C23)
Maximum/minimum operations
(C99)
(C23)    
Exponential functions
(C23)
(C99)
(C99)
(C23)
(C23)
(C99)
(C99)(C23)
(C23)
(C23)
Power functions
(C99)
(C23)
(C23)
(C99)
(C23)
(C23)
Trigonometric and hyperbolic functions
(C23)
(C23)
(C23)
(C23)
(C99)
(C99)
(C99)
Error and gamma functions
(C99)
(C99)
(C99)
(C99)
Nearest integer floating-point operations
(C99)(C99)(C99)
(C99)
(C99)(C99)(C99)
(C23)(C23)(C23)(C23)
Floating-point manipulation functions
(C99)(C99)
(C99)(C23)
(C99)
Narrowing operations
(C23)
(C23)
(C23)
(C23)
(C23)
(C23)
Quantum and quantum exponent functions
Decimal re-encoding functions
Total order and payload functions
Classification
(C99)
(C99)
(C99)
(C23)
Types
(C99)(C99)    

(C99)(C99)    

Macro constants
Special floating-point values
(C99)(C23)
Arguments and return values
FP_NORMALFP_SUBNORMALFP_ZEROFP_INFINITEFP_NAN
(C99)(C99)(C99)(C99)(C99)    
Error handling
 
Defined in header <math.h>
#define FP_NORMAL    /*implementation defined*/
(since C99)
#define FP_SUBNORMAL /*implementation defined*/
(since C99)
#define FP_ZERO      /*implementation defined*/
(since C99)
#define FP_INFINITE  /*implementation defined*/
(since C99)
#define FP_NAN       /*implementation defined*/
(since C99)

The FP_NORMAL, FP_SUBNORMAL, FP_ZERO, FP_INFINITE, FP_NAN macros each represent a distinct category of floating-point numbers. They all expand to an integer constant expression.

Constant Explanation
FP_NORMAL indicates that the value is normal, i.e. not an infinity, subnormal, not-a-number or zero
FP_SUBNORMAL indicates that the value is subnormal
FP_ZERO indicates that the value is positive or negative zero
FP_INFINITE indicates that the value is not representable by the underlying type (positive or negative infinity)
FP_NAN indicates that the value is not-a-number (NaN)

Example

#include <stdio.h>
#include <math.h>
#include <float.h>
 
const char *show_classification(double x) {
    switch(fpclassify(x)) {
        case FP_INFINITE:  return "Inf";
        case FP_NAN:       return "NaN";
        case FP_NORMAL:    return "normal";
        case FP_SUBNORMAL: return "subnormal";
        case FP_ZERO:      return "zero";
        default:           return "unknown";
    }
}
int main(void)
{
    printf("1.0/0.0 is %s\n", show_classification(1/0.0));
    printf("0.0/0.0 is %s\n", show_classification(0.0/0.0));
    printf("DBL_MIN/2 is %s\n", show_classification(DBL_MIN/2));
    printf("-0.0 is %s\n", show_classification(-0.0));
    printf(" 1.0 is %s\n", show_classification(1.0));
}

Output:

1.0/0.0 is Inf
0.0/0.0 is NaN
DBL_MIN/2 is subnormal
-0.0 is zero
 1.0 is normal

References

  • C17 standard (ISO/IEC 9899:2018):
  • 7.12/6 FP_NORMAL, ... (p: 169-170)
  • C11 standard (ISO/IEC 9899:2011):
  • 7.12/6 FP_NORMAL, ... (p: 232)
  • C99 standard (ISO/IEC 9899:1999):
  • 7.12/6 FP_NORMAL, ... (p: 213)

See also

classifies the given floating-point value
(function macro)