ldexp() function in C/C++

The ldexp() function in C computes the result of multiplying a floating-point number by an integral power of 2. It calculates x * 2^exp where x is a floating-point value and exp is an integer exponent.

Syntax

float ldexp(float x, int exp);
double ldexp(double x, int exp);
long double ldexp(long double x, int exp);

Parameters

• x − The floating-point base value
• exp − The integer exponent representing power of 2

Return Value

Returns x * 2^exp. If the result is too large to represent, it returns HUGE_VAL (infinity).

Example 1: Basic Usage

This example demonstrates basic usage of ldexp() function −

#include <stdio.h>
#include <math.h>

int main() {
    double a = 10.0;
    int exp = 2;
    double result = ldexp(a, exp); // Calculates 10 * 2^2 = 10 * 4 = 40
    printf("ldexp(%.1f, %d) = %.1f\n", a, exp, result);
    return 0;
}

ldexp(10.0, 2) = 40.0

Example 2: Overflow Condition

When the result is too large to represent, ldexp() returns infinity −

#include <stdio.h>
#include <math.h>

int main() {
    double a = 10.0;
    int exp = 5000;
    double result = ldexp(a, exp); // Very large exponent causes overflow
    printf("ldexp(%.1f, %d) = %f\n", a, exp, result);
    return 0;
}

ldexp(10.0, 5000) = inf

Example 3: Different Data Types

The ldexp() function works with different floating-point types −

#include <stdio.h>
#include <math.h>

int main() {
    float f = 3.5f;
    double d = 7.25;
    
    printf("ldexp(%.1ff, 3) = %.1f\n", f, ldexp(f, 3));
    printf("ldexp(%.2f, -2) = %.2f\n", d, ldexp(d, -2));
    return 0;
}

ldexp(3.5f, 3) = 28.0
ldexp(7.25, -2) = 1.81

Key Points

• The function header <math.h> must be included to use ldexp().
• Negative exponents divide by powers of 2 instead of multiplying.
• The function is commonly used in floating-point arithmetic and scientific calculations.

Conclusion

The ldexp() function provides an efficient way to multiply floating-point numbers by powers of 2. It handles overflow gracefully by returning infinity when results exceed representable limits.