From: "Rob Kramer" <robk AT cyberway DOT com DOT sg>
Newsgroups: comp.os.msdos.djgpp
References: <3a8ca532$1 AT news DOT starhub DOT net DOT sg> <7458-Fri16Feb2001081657+0200-eliz AT is DOT elta DOT co DOT il> <3a8cd661$1 AT news DOT starhub DOT net DOT sg> <1659-Fri16Feb2001123414+0200-eliz AT is DOT elta DOT co DOT il>
Subject: Re: System clock problem.
Date: Fri, 16 Feb 2001 21:04:40 +0800
Lines: 36
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.50.4133.2400
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
NNTP-Posting-Host: 203.117.157.106
Message-ID: <3a8d25a8$1@news.starhub.net.sg>
X-Trace: 16 Feb 2001 21:05:44 +0800, 203.117.157.106
Organization: StarHub Internet Pte Ltd
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com


"Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il> wrote in message
news:1659-Fri16Feb2001123414+0200-eliz AT is DOT elta DOT co DOT il...
> > From: "Rob Kramer" <robk AT cyberway DOT com DOT sg>
> > Newsgroups: comp.os.msdos.djgpp
> > Date: Fri, 16 Feb 2001 15:18:41 +0800
> >
> > Correct, there's about 50 of these devices (embedded PCs). 3 are
starting to
> > show the problem now. I'm suspecting a hardware problem, but can't
explain
> > why all is fine again after I exit my application.
>
> Based on your description, I won't suspect the hardware first, because
> the system clock is evidently okay once you exit the program.
>
> What other time-related functions or interrupts does that program use?
> Does it delay itself, for example, or use usleep?

Lots of delays and timings yes, but they are all derived from a 5ms Allegro
timer, which reprograms one of the hardware timer/counters. I use the
Allegro timer interrupt to add 5 to a global counter, so I have a
millisecond counter with 5ms precision. All timings are done by using that
counter. No usleep used.

The clock-string written to the screen is directly converted from time(),
using my own conversion routine (strftime() was not suitable for 'historical
reasons'). Going through the code, there's no way a clock speedup could be
introduced here..

Thanks,

    Rob