On Thu, Jun 08, 2006 at 03:21:17PM +0300, Sami Hangaslammi wrote: > I'm not an expert on these issues, but the way I understand it, sleep > is so inaccurate, that for very small delays, you need a use a busy > loop that polls the timer (causing 100% CPU utilization), whereas for > longer delays sleep works ok.
I did a little test and at least on my machine, time.sleep(0.01) was about as good as pygame.time.wait(10). Neither were ever more than 10% off, though sleep was always under, and wait was both under and over. But I also did some reading on framerates and decided to shoot for 40fps instead of 100. That at least brought my cpu usage down to 40-some percent using wait. Perhaps I should stick with pygame's time stuff in the interest of cross-platform consistency. (haven't tested on any other machines yet) -- Nick Welch | mack @ incise.org | http://incise.org
