Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!elroy.jpl.nasa.gov!jato!vsnyder
From: vsnyder@jato.jpl.nasa.gov (Van Snyder)
Newsgroups: comp.sys.atari.st
Subject: Re: (none) Was: Multitasking
Message-ID: <1991Mar8.213858.11770@jato.jpl.nasa.gov>
Date: 8 Mar 91 21:38:58 GMT
References: <74457@bu.edu.bu.edu> <1010002@hpmwmat.HP.COM> <1401@TALOS.UUCP>
Reply-To: vsnyder@jato.Jpl.Nasa.Gov (Van Snyder)
Organization: Jet Propulsion Laboratory, Pasadena, CA
Lines: 36

In article <1401@TALOS.UUCP> jerry@TALOS.UUCP (Jerry Gitomer) writes:
>mikep@hpmwtd.HP.COM (Mike Powell) writes:
>
>		[Definitions of multitasking omitted]
>
>>	A good multitasking system shows little or no signs of slowdown
>>	when multiple tasks are run.  The degree to which this is true
>>	is one good measure of how well a multitasking system it is.
>
>This is more dependent on the nature of the tasks being run than
>on the quality of the multitasking system.  For example if
>running two input/output intensive tasks there will no apparent
>performance degradation, when running one processor intensive
>task and one input/output intensive task there will be little
>apparent performance degradation, but when running two processor
>intensive tasks there will be significant apparent performance degradation.
>
>Since most users of personal computers tend to run one
>interactive task (input/output intensive) when they are
>multitasking there is little apparent performance degradation,
>but if the same user, using the same machine, were to run two
>processor intensive tasks on the same system you would find that
>the elapsed time to perform the two tasks would probably exceed
>the elapsed time to perform the same two tasks sequentially.

An interesting measure of the quality of a multitasking system is how fast
a single application runs, as compared to running that application on a 
non-multitasking system.  The current batch of crud seems to waste about
33% of your cycles.  Since I've seen multitasking systems that waste more
like 4-5% of the cycles when running one application, the 33% wasted by Unix
isn't inevitable.  Where does it go?

-- 
vsnyder@jato.Jpl.Nasa.Gov
ames!elroy!jato!vsnyder
vsnyder@jato.uucp