Multiplication with constant - imul or shl-add-combination

Question

This question is about how we multiply an integer with a constant. So let\'s look at a simple function:

int f(int x) {
    return 10*x;
}

How c

Answer 1

I'd guess that the shift-and-add sequence was faster than imul; this has been true for many versions of x86 chips. I don't know if it is true of Haswell; still, doing a imul in 2 clock cycles takes significant chip resources if it is doable at all.

I'm a bit surprised it didn't produce an even faster sequence:

 lea   y, [2*y]
 lea   y, [5*y]

[OP edits his answer, shows optimized code producing ADD then LEA. Yes, that's a better answer; the ADD r,r is smaller spacewise than lea ..[2*y] so the resulting code is smaller and the same speed]