On Mon, Jun 29, 2009 at 11:42 AM, Andrew Haley<aph@xxxxxxxxxx> wrote: > This assumption is wrong. main() is called by another function, usually > something in crtn, and gcc has no way to know that the calling function > will not access pixels when main() exits. Of course - what happens _after_ main() cannot be anlayzed by the optimizer, because pixels could have been passed around to somewhere else by resize(). But this is also completely irrelevant to optimizing this loop. That was what I was saying: After the call to resize resize(), there is nothing else apart from main itself that can possibly change pixels until main() exits! Oh and by the way: I have tested this in Visual Studio as well and it is able to analyze this situation correctly. There is essentially no difference in the loop-code it generates, no matter whether resize is inlined or not. IMHO gcc is being very clever in this situation.... Cheers Frank
