Count bits used in int

Question

If you have the binary number 10110 how can I get it to return 5? e.g a number that tells how many bits are used? There are some likewise examples listed below:

Answer 1

From here, a way to do it with just bitwise-and and addition:

int GetHighestBitPosition(int value) {
  if (value == 0) return 0;

  int position = 1;
  if ((value & 0xFFFF0000) == 0) position += 16;
  if ((value & 0xFF00FF00) == 0) position += 8;
  if ((value & 0xF0F0F0F0) == 0) position += 4;
  if ((value & 0xCCCCCCCC) == 0) position += 2;
  if ((value & 0xAAAAAAAA) == 0) position += 1;

  return position;
}

Answer 2

You really just want to find the position of the highest bit that is a 1. See this page, under the heading "Finding integer log base 2 of an integer (aka the position of the highest bit set)".

Answer 3

Be careful what you ask for. One very fast technique is to do a table lookup:

int bittable [] = {0, 1, 1, 2, 1, 2, 2, 3, 1, 2, ... };
int numbits (int v)
{
    return bittable [v];
}

where bittable contains an entry for every int. Of course that has complications for negative values. A more practical way would be to count the bits in bitfields of the number

int bittable [16] = {0, 1, 1, 2,  1, 2, 2, 3,  1, 2, 2, 3,  2, 3, 3, 4};
int numbits (int v)
{
    int s = 0;
    while (v != 0)
    {
         s += bittable [v & 15];
         v >>= 4;
    }
    return s;
}

Answer 4

int CountBits(uint value)
{
    for (byte i = 32; i > 0; i--)
    {
        var b = (uint)1 << (i - 1);
        if ((value & b) == b)
            return i;
    }
    return 0;
}

Answer 5

Easiest?

32 - Integer.numberOfLeadingZeros(value)

If you are looking for algorithms, the implementors of the Java API agree with your divide-and-conquer bitshifting approach:

public static int numberOfLeadingZeros(int i) {
    if (i == 0)
        return 32;
    int n = 1;
    if (i >>> 16 == 0) { n += 16; i <<= 16; }
    if (i >>> 24 == 0) { n +=  8; i <<=  8; }
    if (i >>> 28 == 0) { n +=  4; i <<=  4; }
    if (i >>> 30 == 0) { n +=  2; i <<=  2; }
    n -= i >>> 31;
    return n;
}

Edit: As a reminder to those who trust in the accuracy of floating point calculations, run the following test harness:

public static void main(String[] args) {
    for (int i = 0; i < 64; i++) {
        long x = 1L << i;
        check(x);
        check(x-1);
    }
}

static void check(long x) {
    int correct = 64 - Long.numberOfLeadingZeros(x);
    int floated = (int) (1 + Math.floor(Math.log(x) / Math.log(2)));
    if (floated != correct) {
        System.out.println(Long.toString(x, 16) + " " + correct + " " + floated);
    }
}

The first detected deviation is:

ffffffffffff 48 49

Answer 6

Unfortunately there is no Integer.bitLength() method that would give you the answer directly.

An analogous method exists for BigInteger, so you could use that one:

BigInteger.valueOf(value).bitLength()

Constructing the BigInteger object will make it somewhat less efficient, but that will only matter if you do it many millions of times.