Re: Problems with fast timeouts



On Sat, 11 Jan 2003 13:53:27 +0100, Krzysztof Dubowik said:
> It's a simulator of a measuring device where the sampling rate is fixed at
> 200Hz. I want to draw the graph at "real-time" - 1 pixel per 5ms.

Optimization 1: Remember that the actual monitor probably won't be refreshing
any faster than 75Hz, so you can batch up updates and only push them out once
per refresh.

Optimization 2: Remember that monitors refresh at 72/75Hz so their refresh
rate is far above the "flicker" rate, and that most of the time, you can
get away with MUCH lower refresh rates - most of the time, people don't
complain in the slightest about movies at 24fps or TV at 30fps.

You may be able to get a BIG performance boost if you run the simulation
at an effective 200hz, but batch up your updates 8 at a time at 25fps.  Or
you might look at using a 40ms timeout and doing 8 per timeout - this has
the advantage of being a long enough timeout that most kernels will be able
to deal with it in a sane manner....
-- 
				Valdis Kletnieks
				Computer Systems Senior Engineer
				Virginia Tech

Attachment: pgptDvOmRbp2j.pgp
Description: PGP signature



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]