Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!dali.cs.montana.edu!uakari.primate.wisc.edu!zaphod.mps.ohio-state.edu!rpi!uupsi!rodan.acs.syr.edu!isr
From: isr@rodan.acs.syr.edu (Michael S. Schechter - ISR group account)
Newsgroups: comp.realtime
Subject: Re: In search of Real-Time OS for 386 box
Message-ID: <1990Aug27.145954.26525@rodan.acs.syr.edu>
Date: 27 Aug 90 14:59:54 GMT
References: <7629@helios.TAMU.EDU>
Organization: Institute for Sensory Research
Lines: 24

In article <7629@helios.TAMU.EDU> hyeom@cs.tamu.edu (Heon Y Yeom) writes:
>My current approach to do this is using the timer interrupt to interrupt
>the CPU 10,000 times per second which doesn't work too good.
>Heon Yeom
>hyeom@photon.tamu.edu
>hyy5325@tamvenus


I would think that to get 10k/sec task scheduling you'd have to
use an interrupt for it.. that's only 100 usec between them and
from a lot of the descrptions of the OS's posted it doesn't seem
like their anywhere near that fast. (Mind you, I haven't used any
real-time OS's since RT11, i can get away with simply disabling all
the 'normal' spuriosu OS interrupts inmy stuff)
But, by using interrupts you shoudn't have any problem acheiving that
speed on a 386. I've gotten a 5k/sec task to run on a original 8Mhz
IBM brand AT. (admittedly it was a __very__ short task).
of course, it will mess up timekeeping unless you keep track of the time
you had spent with your timer task in place of the system timer task
and then fix the system's clock to the correct time..

--
Mike Schechter, Computer Engineer,Institute Sensory Research, Syracuse Univ.
InterNet: Mike_Schechter@isr.syr.edu isr@rodan.syr.edu Bitnet: SENSORY@SUNRISE