Date: Mon, 29 Sep 2003 09:11:51 -0600 From: "Kenneth D. Merry" <ken@kdm.org> To: Edy Lie <email@edylie.net> Cc: freebsd-scsi@freebsd.org Subject: Re: Benchmark Raid 5 SCSI U160 Message-ID: <20030929151151.GA5603@panzer.kdm.org> In-Reply-To: <1064847130.604.58.camel@ps2.consoledojo.com>
index | next in thread | previous in thread | raw e-mail
On Mon, Sep 29, 2003 at 22:52:10 +0800, Edy Lie wrote: > Greetings, > > I have done a benchmark on the following configuration. > > RAID 5 SCSI U160 Benchmark > > The following is the benchmark result > OS: FreeBSD 4.8p4 > Dual 1GHZ Proc > 1GB ECC RAM > 18GBx5 (RAID5) 10K RPM > Adaptec 2100s (32MB ECCRAM) > Software: bonnie > parameter: bonnie -m ANUBIS -s 100M > > -------Sequential Output-------- ---Sequential Input-- --Random-- > -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks--- > Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU > /sec%CPU > > ANUBIS 100 3959 7.9 4065 2.8 3763 2.5 33936 100.0 497807 100.0 39277.7 > 198.5 > > I do not have other to compare. what do you guys think about the result > ? I assume you're going through the filesystem? If so, you need to make sure that your test file size is a good bit larger than your memory size. 100MB won't quite do the trick. You need something like a 4GB file to make sure you aren't really getting any caching effects, especially on reads. You can tell you're hitting the cache on the block sequential reads. A single channel Ultra160 RAID controller can't do 498MB/sec. Ken -- Kenneth Merry ken@kdm.orghome | help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030929151151.GA5603>