Optimization suggestion

AND51 · Post by **AND51** » Mon Feb 25, 2008 3:18 pm

pdwyer wrote:Interestingly I get much closer results when I swap the position of the functions around.

Ah, I'm not the onliest one noticing this problem...
This occured already at other performance tests on my machine, too...
In some cases, the first loop was always a bit slowier than the following loops.

However, the inaccuracy with ElapsedMilliseconds() is IMHO not significant, if the difference is big enough:
(141-16)-(62+16) = 47 ms
47 is still indicates that the second loops is faster, even if ElapsedMilliseconds() misses the actual time by 16 ms.

pdwyer · Post by **pdwyer** » Mon Feb 25, 2008 3:23 pm

true, it will skew a % figure though. I just made the 3 a 10 to make it a little clearer

Psychophanta · Post by **Psychophanta** » Sat Feb 07, 2009 11:25 am

I don't use C so then i have nothing related to C installed in my pcs.
But Can someone please tell us how the hell does it the best C/C++ compiler on the world, or for example MS VC++

Trond · Post by **Trond** » Sat Feb 07, 2009 11:29 am

Gcc uses the mov variant.

Psychophanta · Post by **Psychophanta** » Sat Feb 07, 2009 11:31 am

Trond wrote:Gcc uses the mov variant.

Mmmhhh... yes,
but i guess Gcc is far to be the best or even between the bests...

Thinking twice, take in account that the times you plainly assign a variable to another one is insignificant compared to the times you assign a expression to a variable in any piece of code or in any big app.

blueznl · Post by **blueznl** » Sat Feb 07, 2009 11:35 am

Makes no difference whatsoever on my machine.

gnasen · Post by **gnasen** » Sat Feb 07, 2009 2:45 pm

nearly every time the same result

109 vs 109