xfs
[Top] [All Lists]

Re: benchmarks

To: "P.Dixon" <P.Dixon@xxxxxxxxx>
Subject: Re: benchmarks
From: Eric Sandeen <sandeen@xxxxxxx>
Date: Fri, 13 Jul 2001 09:23:19 -0500
Cc: linux-xfs@xxxxxxxxxxx
References: <Pine.LNX.4.33.0107130921430.29969-100000@heppct.ph.qmw.ac.uk>
Sender: owner-linux-xfs@xxxxxxxxxxx
"P.Dixon" wrote:
> 
> Hello,
> 
> Any idea why xfs appears to be very much slower than reiserfs with these
> benchmarks:
> 
> http://www.namesys.com/benchmarks/mongo/2.4.5-xfs-ext2_vs_reiserfs.html
> 
> Admittedly, the benchmarks were done by namesys...

More than _done_ by namesys, but _written_ by namesys.

[eric@stout eric]$ tar xvzf mongo.tar.gz 
[eric@stout eric]$ cd mongo_pl
[eric@stout mongo_pl]$ more mongo.pl

#!/usr/bin/perl 
#
# Copyright 2000 by Hans Reiser, licensing governed by reiserfs/README
#

#
# Mongo.pl is reiserfs benchmark. 
#
...

I'm not saying that there's anything sinister going on here (really!)
but when you have a benchmark suite written by the filesystem author,
it's probably not the most objective test of other filesystems.  I don't
know the history of mongo, but I'd be willing to guess that it's
designed to test reiserfs performance, to be used in reiserfs testing,
and it may not be the best thing to compare various filesystems.

Anybody want to do some independent iozone (or other benchmark) testing?

-Eric


<Prev in Thread] Current Thread [Next in Thread>