Strangely, I can't reproduce the effect with a divide by 16. With GCC 4.5.2 -O2, I see a "shrl $4, %eax" for unsigned, and "sarl $4, %eax" for signed. However, if I divide by 2, the results are different; the function:

  int div_by_2(int x) {return x/2;}

compiles to:

  div_by_2:
    movl	%edi, %eax
    shrl	$31, %eax
    addl	%edi, %eax
    sarl	%eax
    ret

Meanwhile, the signed variant:

  unsigned int div_by_2(unsigned int x) {return x/2;}

compiles to:

  div_by_2:
    movl	%edi, %eax
    shrl	%eax
    ret

My conclusion remains that for hot loops, you should check the compilers assembly to make sure the compiler is applying the strength reduction you expect.