Hacker News Reader

Comment by taeric | Hacker News Reader

1 isn't really the same thing. I don't know any compiler that would turn

    for (i = 0; i < length; i++)

into

    for (i = length; i != 0; i--)

Which is what I read 1 to be.

As for why you sometimes do one over the other. I think it is true that computers branch on zero more easily than other values. That said, I'd be a little surprised to know it makes a difference that can be measured in any meaningful program.

View on HN

Clang will if `i` is not referenced anywhere else. e.g.

    int i,n;
    for (i = 0; i < length; i++)
        n = something();
    return n;

It's also free to treat `i` as a different variable:

    int i,n = 0;
    for (i = 0; i < length; i++)
        something(n++); /* Compiler will merge the n and i variables */
    return n;

Or eliminate the loop completely if there are no side effects.

    int i, n = 0;
    for (i = 0; i < length; i++)
        n = i + 1; /* A very roundabout way to say "n = length" */
    return n;

In neither of those did the sequence change direction, though?

See this is why I felt I was obsessing over something very small.

Loop inversion, unrolling, and splitting. The compiler can do so much with a loop to make it run faster on 'this machine' to optimize for speed, memory use, or binary size.

Sorry, I didn't copy the assembler output for brevity but Clang generates `decl %esi` (or `addl $-1,%ebx` depending on the version) in the first case.

Do you know of any cases where it changes the sequence of seen values? I'm assuming it can in this case, because the i is never used?

Reply on news.ycombinator.com