Path: utzoo!attcan!uunet!cs.utexas.edu!mailrus!sharkey!inquiry!shadooby!egrunix!nucleus!dar
From: dar@nucleus.UUCP (Dario Alcocer)
Newsgroups: comp.sys.ibm.pc
Subject: Re: getting the time
Summary: BIOS clock ticks & time
Message-ID: <5487@nucleus.UUCP>
Date: 20 Nov 89 18:29:36 GMT
References: <17118@netnews.upenn.edu>
Organization: The Nucleus Public Access Unix, Clarkston, MI
Lines: 33

In article <17118@netnews.upenn.edu>, hardt@linc.cis.upenn.edu (Dan Hardt) writes:
> A friend needs to read the time for a real-time application.
> We tried using some Microsoft C functions for reading the
> clock.  The functions returned the time down to the hundredth
> of a second -- but they don't seem to notice any change in the
> time less than 5/100 of a second or so.  If we have a loop which
> continually checks the time, its values are ,eg, 
>         0,0,0,0,5,5,5,5,5,11,11,11,etc
> Why is this?  Any way to get a more fine-grained access to the time?

If i'm not mistaken, it seems that the functions you are using
are getting the BIOS clock ticks since midnight (this value is
used in INT 21h, 2Ch: DOS Get System Time).  The BIOS clock tick
count is available at 0040:006C.

Since the BIOS clock tick count is incremented by INT 8, which is called
approximately 18 times per second, the granularity of the clock tick
count is ~1/18 sec, which is roughly equal to 5/100 sec.  To have smaller
granularity would require you to reprogram the timer chip, and I have seen
it done before, although I have no code that does this.  Maybe someone
else has done it, and could share this with you.


Hasta...

dario.
    
+==================================================================+
|                               +                                  |
| Dario Alcocer (San Diego, CA) | Internet......dar@nucleus.mi.org |
| via Nucleus                   | phone...............619-450-1667 |
|                               +                                  |
+==================================================================+