Date: Sat, 8 Nov 1997 11:45:21 +0100 From: Stefan Esser <se@FreeBSD.ORG> To: Amancio Hasty <hasty@rah.star-gate.com> Cc: hackers@FreeBSD.ORG Subject: Re: Seagate Cheetah performance... Message-ID: <19971108114521.06151@mi.uni-koeln.de> In-Reply-To: <199711080626.WAA08424@rah.star-gate.com>; from Amancio Hasty on Fri, Nov 07, 1997 at 10:26:34PM -0800 References: <199711080104.RAA00593@rah.star-gate.com> <199711080626.WAA08424@rah.star-gate.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 1997-11-07 22:26 -0800, Amancio Hasty <hasty@rah.star-gate.com> wrote:
> Well,
>
> I found a way to nearly double my performance:
>
>
> iozone 48
>
> IOZONE: Performance Test of Sequential File I/O -- V2.01 (10/21/94)
> By Bill Norcott
>
> Operating System: FreeBSD 2.x -- using fsync()
>
> Send comments to: b_norcott@xway.com
>
> IOZONE writes a 48 Megabyte sequential file consisting of
> 98304 records which are each 512 bytes in length.
^^^^^^^^^ This is where
your problem originates!
Reading 512 byte blocks is not a typical situation at all ...
> It then reads the file. It prints the bytes-per-second
> rate at which the computer can read and write files.
>
>
> Writing the 48 Megabyte file, 'iozone.tmp'...5.523438 seconds
> Reading the file...4.921875 seconds
>
> IOZONE performance measurements:
> 9112377 bytes/second for writing the file
> 10226112 bytes/second for reading the file
This is > 20000 SCSI commands ! SCSI overhead must be lower than 50us !
Wow! What are you complaining about ? :-)
> > Any tweaking that I do to improve the performance of my :
> > <SEAGATE ST34501W 0017> type 0 fixed SCSI 2
> >
> > iozone 48
> > IOZONE performance measurements:
> > 5418377 bytes/second for writing the file
> > 7828008 bytes/second for reading the file
Again: You are using a much to small transfer rate !
> > My Ultra DMA IDE disk generates about 8MB/sec for writing .
IDE overhead is much smaller than SCSI, there is no arbitration
for a shared bus, for example.
Please test all three drives with Bonnie. Use a test file at
least twice as much as your system's RAM, and report your
findings ...
You'll get a different picture, I assume !
Regards, STefan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19971108114521.06151>