convert floating point number to fixed point number

float2fix.c

#include <stdio.h>
#include <math.h>
#define FIXED_BIT 12

unsigned short int float2fix(float n)
{
    unsigned short int int_part = 0, frac_part = 0;
    int i;
    float t;

    int_part = (int)floor(n) << FIXED_BIT;
    n -= (int)floor(n);

    t = 0.5;
    for (i = 0; i < FIXED_BIT; i++) {
        if ((n - t) >= 0) {
            n -= t;
            frac_part += (1 << (FIXED_BIT - 1 - i));
        }
        t = t /2;
    }

    return int_part + frac_part;
}

int main()
{
    float n;
    n = 2.5; // 0d10240
    printf("%f => %d\n", n, float2fix(n));
    n = 2.625; // 0d10752
    printf("%f => %d\n", n, float2fix(n));
    n = 0.375; // 0d1536
    printf("%f => %d\n", n, float2fix(n));
    return 0;
}

@sponthus looks like they made that change sometime in the last 5 years since I posted. Here’s a working copy on the wayback machine

Thanx that's perfect !