Path: utzoo!attcan!ncrcan!becker!censor!comspec!tvcent!lethe!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!ucsd!helios.ee.lbl.gov!hellgate.utah.edu!cc.utah.edu!cc.usu.edu!jrd
From: JRD@cc.usu.edu (Joe Doupnik)
Newsgroups: comp.sys.ibm.pc.programmer
Subject: Re: timing a software function via timer chip
Message-ID: <23656@cc.usu.edu>
Date: 5 May 90 03:12:18 GMT
Lines: 16
X-Local-Date: 4 May 90 20:12:18 PDT


	If one takes a careful look at the IBM PC specs Timer channels 0, 1,
and 2 are nearly identical, but channels 0 and 1 are preallocated to memory
refresh and time of day. So, channel 2 is the only one free for ordinary
programming use. Since they are the same kind, driven from the same clock
one needs to simply choose a nice divisor for the chip to count in units
of up to 1/18.2 seconds.
	The chip can be started just before your test function is executed
and then then stopped at the end. The differences in counts before and
after gives the elapsed time in tics; arithmetic yields the time in seconds.
	If you would like to see an example of this in action take a look
at file MSXIBM.ASM in MS Kermit where I calibrate the duration of a
software timing loop to be 1/1000 second by using channel 2 above to measure
the real time of the loop. That file is available via anonymous ftp from
watsun.cc.columbia.edu  in directory kermit/a.
	Joe D.