C++ loop unfolding, bounds

Question

I have a loop that I want to unfold:

for(int i = 0; i < N; i++)
    do_stuff_for(i);

Unfolded:

for(int i = 0; i < N; i += CHUNK) {
    do_stuff_for(i + 0);
    do_stuff_for(i + 1);
    ...
    do_stuff_for(i + CHUNK-1);
}

But, I should make sure that I do not run out of the original N, like when N == 14 and CHUNK == 10. My question is: what is the best/fasters/standard/most elegant (you name it) way to do it?

One solution that comes is:

int i;
for(i = 0; i < (N % CHUNK); i++) 
    do_stuff_for(i);

for(i; i < N; i += CHUNK) {
    // unfolded, for the rest
}

But maybe there is a better practice

@Tomalak Geret'kal: Visual C++ 10 will usually not attempt such aggressive optimizations. — sharptooth, Nov 15 '11 at 12:19
a templated solution? that makes it easy to disable it for modern(*grin*) compilers.. — Karoly Horvath, Nov 15 '11 at 12:21
@TomalakGeret'kal: I noticed that when I do it myself, I get performance gain. — Jakub M., Nov 15 '11 at 12:23
The "best way" to unroll is to let the compiler do it. It will know the optimum number of iterations inside and automatically generate the last non-even iterations as well. This screams micro-optimization that will cost you much more in maintenance than it pays you. — Mark B, Nov 15 '11 at 14:07

score 7 · Answer 1 · answered Nov 15 '11 at 12:18

7

You could use a switch-case.

It's called Duff's Device

answered Nov 15 '11 at 12:18

Yochai Timmer

48,127
24
147
185

I am impressed with this D.D.! – Jakub M. Nov 15 '11 at 12:23
1

@yi_H: indeed, fortunately it's a well-known little optimization trick, and normally obsolete now (though it might sometimes be needed in low-level routines). – Matthieu M. Nov 15 '11 at 12:36

C++ loop unfolding, bounds

1 Answers1

Linked