Question

. e.g. 0xFE10, should output 0xF(1111 in binary).

This is a Qualcomm interview question. This is my idea so far:

I am calling the 16-bit integer:

int num = /*whatever the number is*/

Have four bit masks:

int zeroTo4 = (num & 0x000F);  
int fiveTo5 = (num & 0x00F0) >> 4;   
int eightTo12 = (num & 0x0F00) >> 8;   
int twelveTo16 = (num & 0xF000) >> 12;   

int printbit = zeroTo4;

if( fiveTo5 > printbit )
    printbit = fiveTo5;

if( eightTo12 > printbit )
    printbit = eightTo12;

if( twelveTo16 > printbit )
    printbit = twelveTo16;

printf( "Largest bit of %X is %1X\n", num, printbit );

However, I’m pretty sure there’s a simpler and easier way to do this. Can anyone help me out? Thanks!

Answer 1

int max4(int j)
{
   int ret=0;
   while(j>0)
   {
        if( (j&0xf) > ret ) ret=j&0xf;
        j>>=4;
   }
   return ret;
}

Some may prefer:

int max4(int j)
{
   int ret=0;
   do if( (j&0xf) > ret ) ret=j&0xf;
   while((j>>=4)>0);
   return ret;
}