Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!uunet!bu.edu!snorkelwacker.mit.edu!thunder.mcrcim.mcgill.edu!poly-vlsi!big
From: big@vlsi.polymtl.ca (Patrick Drolet)
Newsgroups: comp.os.msdos.programmer
Subject: Re: Accurate timing request
Message-ID: <1991Mar10.180448.20645@vlsi.polymtl.ca>
Date: 10 Mar 91 18:04:48 GMT
References: <1991Mar6.145414.8747@uwovax.uwo.ca>
Sender: news@vlsi.polymtl.ca (USENET News System)
Organization: Ecole Polytechnique de Montreal
Lines: 25

In article <1991Mar6.145414.8747@uwovax.uwo.ca> bangarth@uwovax.uwo.ca writes:
>Researchers here are looking for some facility to measure time with 1 ms. 
>or less accuracy on an AT or 386 machine. We currently with our software 
>(in Turbo Pascal) can't get any better than about 3 ms. accuracy. I have heard
>that there was an article written in a recent tech magazine describing a way
>someone has gotten the required accuracy. I know this is vague, but does anyone
>know where this is written up, or have a more direct answer? Thanks for your
>help. It could save me a lot of library work.

I'm afraid I only have half of the answer (since I don't have my Pheonix BIOS
tehc ref near by).  There's an address (in low RAM) that instruct the real-time
clock to generate an interrupt every X milliseconds...  You could take the
actual 55ms to 1 ms, BUT BE CAREFULL!!!  Be sure that no resident software are
hooked on the timer interrupt, and make sure you're using a FAST computer, sincethe overhead of the interrupt, for this purpose, is about .3 and a 386/20

Also remember to tell your prog to reset the timer period to 55 ms after use...


							Patrick Drolet
							big@info.polymtl.ca

P.S.: If you wish to be even more accurate, grad the timer interrupt, but
      don't chain with the old interrupt after use!  The only problem is
      that your real -time might not be good...  Bah!  As long as it does not
      affect CMOS ram, all you have to do is to reboot the computer.