From:	Tom Burgess <Tom_Burgess AT bc DOT sympatico DOT ca>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: testing uclock()
Date:	Fri, 25 Apr 1997 21:15:50 -0700
Organization:	BCTEL Advanced Communications
Lines:	18
Message-ID:	<33618176.377@bc.sympatico.ca>
References:	<1 DOT 5 DOT 4 DOT 32 DOT 19970424124818 DOT 002d7ec8 AT ubeclu DOT unibe DOT ch>
NNTP-Posting-Host:	pntn02m01-36.bctel.ca
Mime-Version:	1.0
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

Roger Noss wrote:
> 
> I am trying to follow the many recommendations to use uclock() but am
> getting inconsistent results.  I am working in a DOS window of WinNT3.51 on
> a 200MHz Pentium Pro.

Don't have the answer, but a few observations:

	- printf of uninitialized t0 is puzzling
	- use of short ints on Pentiums is expensive (more CPU cycles)
	- Since uclock messes with the timer hardware, I am surprised NT
	  doesn't barf. Uclock has no error return code. Maybe uclock
	  just fails on NT and is returning garbage?
	- for best timing, should make sure dt[] is in cache before
	  using it to record results, otherwise cache misses will
	  confuse timing. Just clear it to zero or something.

	regards, tom

- Raw text -