*BSD News Article 28219


Return to BSD News archive

Newsgroups: comp.os.386bsd.questions
Path: sserve!newshost.anu.edu.au!munnari.oz.au!ihnp4.ucsd.edu!news.cerf.net!mvb.saic.com!MathWorks.Com!europa.eng.gtefsd.com!howland.reston.ans.net!agate!boulder!cnsnews!benji.Colorado.EDU!frechett
From: frechett@benji.Colorado.EDU (-=Runaway Daemon=-)
Subject: Re: linux's I/O calls faster than NetBSD's ?
Message-ID: <CMFvuI.IHu@cnsnews.Colorado.EDU>
Sender: usenet@cnsnews.Colorado.EDU (Net News Administrator)
Nntp-Posting-Host: benji.colorado.edu
Organization: University of Colorado, Boulder
References: <2lhv9r$pbt@homea.ensta.fr>
Date: Thu, 10 Mar 1994 08:06:17 GMT
Lines: 28



Sounds like you managed to get the whole operation into linux's
disk cache.  A similiar effect on bsd can be had if you run
your very disk IO intensive stuff from an MFS drive.   If I run
your program from /tmp where it is basically on a 10M RAM drive

DEBUT
      5000
C'EST TERMINE............
 TEMPS ECOULE EN SECONDES: 16

And from /usr

DEBUT
      5000
C'EST TERMINE............
 TEMPS ECOULE EN SECONDES: 331

Can't quite get it down to 4 seconds but the effect is quite obvious
for anyone who has lot of memory.  If the linux box had less memory
I suspect it would perform closer to *BSD's non-MFS numbers.

I'm not sure if I like unified caches yet as I've seen them seriously
cripple an Alpha box.  Having not spent much time in Linux I can't
say if it's better than OSF/1's implementation or not.

	ian