Question

I’m writing a function, which determine the number of useful bits of a 16 bits integer.

int16_t
f(int16_t x)
{
    /* ... */
}

For example, the number “00000010 00100101” has 10 useful bits. I think I should use some bitwise operators, but I don’t know how. I’m looking for some ways to do it.

Answer 1

If you’re using gcc (or a gcc-compatible compiler such as ICC) then you can use built in intrinsics, e.g.

#include <limits.h>

int f(int16_t x)
{
    return x != 0 ? sizeof(x) * CHAR_BIT - __builtin_clz(x) : 0;
}

This assumes you just want the number of bits to the right of the last leading zero bit.

For MSVC you can use _BitScanReverse with some adjustment.

Otherwise if you need this to be portable then you can implement your own general purpose clz function, see e.g. http://en.wikipedia.org/wiki/Find_first_set