*BSD News Article 39734


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!bunyip.cc.uq.oz.au!munnari.oz.au!spool.mu.edu!howland.reston.ans.net!pipex!uunet!jaring.my!oasys!othman
From: othman@oasys.pc.my (Othman bin Hj. Ahmad)
Newsgroups: comp.os.386bsd.misc
Subject: Re: bonnie on several free unices
Message-ID: <gRZuyc3w165w@oasys.pc.my>
Date: Sun, 11 Dec 94 00:50:51 GMT
References: <rcarterD0rBru.2LH@netcom.com>
Distribution: world
Organization: Personal,Kota Kinabalu,Sabah,Malaysia
Lines: 43

rcarter@netcom.com (Russell Carter) writes:

> In article <Xs6Hyc1w165w@oasys.pc.my>,
> Othman bin Hj. Ahmad <othman@oasys.pc.my> wrote:
> >rcarter@netcom.com (Russell Carter) writes:
> >
...
> 
> Sure.  The file system block sizes are the defaults each OS gets
> "caught" with.  If your interest level is high enough, you could 
I'd surely love to if only I have access.

> look it up.  (Mine isn't.)
> 
> The bonnie source code was of course the same for each test, and
> I can email it to you if you like.

I've bonnie source codes.

> 
> The only benchmarks that are ever worth a damn are your own 
> applications, but I thought these results were kind of

Judging from incessant arguments here, it surely does not tell us much.

> interesting, if a trifle variable.

Given enough information about the benchmarks, we can surely estimate
how applications behave.

Block sizes are very important in determining i/o speeds. In BSD you
can set it up as high as you'd like it to be. You can do other fine
tuning as well. The speed difference is almost equal to the
difference in block sizes, if I can remember correctly.

The default sizes are actually different for different versions and
people  so we cannot rely on them too much.

Disclaimer: I only speak for myself.

SABAH is Heaven.
Perpetual Clear Blue Sky, Sea and Mountains near modern civilisations,
but how long will this last?