In C, frexp() splits x into a fraction and a power of two. The call returns the fraction, usually from 0.5 up to but not including 1, and writes the exponent through a pointer. frexp(8, &exp) returns 0.5 and sets exp to 4, because 0.5 × 24 is 8.
You pass the double and a pointer to an int. Both are required. Include <math.h>.
Pass the value and the address of an int. Start with these lines.
int exp;
double frac = frexp(8.0, &exp);| Call | What it does |
|---|---|
frexp(8.0, &exp) | Returns 0.5 and stores 4 in exp. |
Include <math.h>. On GCC, link with -lm. The pointer must point to an int you own.
double frexp(double x, int *exp);
float frexpf(float x, int *exp);
long double frexpl(long double x, int *exp);| Parameter | Description |
|---|---|
| x | Required. The double to split. |
| exp | Required. A pointer to an int. The function writes the power of two there. |
| Return Value | Description |
|---|---|
| Fraction | The significand. For a normal number it is from 0.5 up to, but not including, 1, and it keeps the sign of x. |
Call frexp and print both the mantissa and the exponent.
#include <stdio.h>
#include <math.h>
int main(void) {
int exp;
double frac = frexp(8.0, &exp);
printf("frac = %.1f\n", frac);
printf("exp = %d\n", exp);
return 0;
}int receives the binary exponent.&exp stores 4.0.5 sits at the low end of [0.5, 1).8 = 0.5 × 2⁴.ldexpldexp is the inverse: multiply the mantissa by a power of two.
#include <stdio.h>
#include <math.h>
int main(void) {
int exp;
double x = 123.45;
double frac = frexp(x, &exp);
double back = ldexp(frac, exp);
printf("frac = %.6f, exp = %d\n", frac, exp);
printf("back = %.2f\n", back);
return 0;
}frexp finds frac and exp = 7 (2⁷ = 128).ldexp(frac, 7) multiplies by 128.frexp/ldexp together when scaling by powers of two.Zero is a special case. Negative inputs keep a negative mantissa with the same exponent magnitude rule.
#include <stdio.h>
#include <math.h>
int main(void) {
int e0, e1;
double f0 = frexp(0.0, &e0);
double f1 = frexp(-4.0, &e1);
printf("0 → frac=%.1f exp=%d\n", f0, e0);
printf("-4 → frac=%.1f exp=%d\n", f1, e1);
return 0;
}-4 = -0.5 × 2³, so the mantissa holds the sign.[0.5, 1).int * for the exponent.Compute √(x² + y²) without intermediate overflow.
9 people found this page helpful