Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!spool.mu.edu!nigel.msen.com!caen!usenet.coe.montana.edu!osyjm
From: osyjm@cs.montana.edu (Jaye Mathisen)
Newsgroups: comp.os.386bsd.misc
Subject: Re: NetBSD IO performance
Date: 3 Sep 1993 04:30:54 GMT
Organization: Computer Science, MSU, Bozeman MT, 59717
Lines: 36
Message-ID: <266h9u$avt@pdq.coe.montana.edu>
References: <9309030243.AA03945@moose.usmcs.maine.edu>
NNTP-Posting-Host: pdq.coe.montana.edu
In article <9309030243.AA03945@moose.usmcs.maine.edu>,
George P. Swanton <swanton@moose.usmcs.maine.edu> wrote:
>for reference, a Decstation 5000/120 (Ultrix) w/ SomeFlavor Seagate SCSI 1G
>
>IOZONE performance measurements:
> 616478 bytes/second for writing the file
> 1001625 bytes/second for reading the file
That's pretty abysmal, unless you're running a version of ultrix before
4.3. I get: (on a 5000/33)
(10 MB file, no chance it was all in the bufcache, as it's only 2 MB's).
IOZONE performance measurements:
1643824 bytes/second for writing the file
2302176 bytes/second for reading the file
under ultrix 4.3A to an HP disk. But that's after going through all
the fs tuning stuff.
But I digress.
>IOZONE performance measurements:
> 187301 bytes/second for writing the file
> 683358 bytes/second for reading the file
Sounds like you need to play with rotdelay, block/frag sizes, and tunefs
a bit.
Nate has a similar config, and to an RZ26 (DEC 1.05GB drive), he was getting
near MB/s writes, 1-1.5MB reads I think. (Memory is fading fast, it was
a long time ago).
--
Jaye Mathisen, COE Systems Manager (406) 994-4780
410 Roberts Hall,Dept. of Computer Science
Montana State University,Bozeman MT 59717 osyjm@cs.montana.edu