*BSD News Article 49638


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!simtel!zombie.ncsc.mil!paladin.american.edu!howland.reston.ans.net!Germany.EU.net!zib-berlin.de!news.tu-chemnitz.de!irz401!uriah.heep!bonnie.heep!not-for-mail
From: j@bonnie.heep.sax.de (J Wunsch)
Newsgroups: comp.periphs.scsi,comp.os.linux.hardware,comp.unix.bsd.freebsd.misc,comp.os.linux.setup
Subject: Re: SCSI PCI host adapter
Date: 29 Aug 1995 09:38:02 +0200
Organization: Private U**x site, Dresden.
Lines: 55
Message-ID: <41ug4q$1vr@bonnie.tcd-dresden.de>
References: <418r3m$9c6@trauma.rn.com> <41eadg$1ajc@news.mindspring.com> <41rqk7$2oj@mailhub.hcl.com> <41t8k9$p3m@news.rrz.uni-koeln.de>
Reply-To: joerg_wunsch@uriah.heep.sax.de
NNTP-Posting-Host: 192.109.108.139
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
Xref: euryale.cc.adfa.oz.au comp.periphs.scsi:36136 comp.os.linux.hardware:14044 comp.unix.bsd.freebsd.misc:4997 comp.os.linux.setup:17953

Stefan Esser <se@MI.Uni-Koeln.DE> wrote:

>The NCR keeps up with any current drive easily,
>and with two drives you'll see some 8MB/s total
>throughput through the file system (as reported
>by 'Bonnie').

Bonnie runs through the file system, right?

Both disks here are Seagate Hawk, running an early version of the
NCR driver on 1.1.5.1 (i586/90, 53C810, 64 MB RAM):

j@bonnie 66% dd if=/dev/rsd0c of=/dev/null bs=64k count=1600
1600+0 records in
1600+0 records out
104857600 bytes transferred in 26 secs (4032984 bytes/sec)
j@bonnie 67% dd if=/dev/rsd0c of=/dev/null bs=64k count=1600 & \
dd if=/dev/rsd1c of=/dev/null bs=64k count=1600
[1] 1954
1600+0 records in
1600+0 records out
104857600 bytes transferred in 28 secs (3744914 bytes/sec)
1600+0 records in
1600+0 records out
104857600 bytes transferred in 28 secs (3744914 bytes/sec)

As one can see, the disk itself ain't the fastest at all, but sucking
from two disks concurrently does slow down the transfers by less than
10 %.

To make this little test complete, i've included the third disk (a lame
Fujitsu one):

j@bonnie 76% dd if=/dev/rsd0c of=/dev/null bs=64k count=1600 & \
dd if=/dev/rsd1c of=/dev/null bs=64k count=1600 & \
dd if=/dev/rsd2c of=/dev/null bs=64k count=1600 
[1] 2031
[2] 2032
1600+0 records in
1600+0 records out
104857600 bytes transferred in 48 secs (2184533 bytes/sec)
1600+0 records in
1600+0 records out
104857600 bytes transferred in 48 secs (2184533 bytes/sec)
1600+0 records in
1600+0 records out
104857600 bytes transferred in 50 secs (2097152 bytes/sec)

This finally thrashes the SCSI bus, yielding only 6.5 MB/s, but not
too bad for a fully-loaded bus compared to the 7.5 MB/s above.
-- 
cheers, J"org                      private:   joerg_wunsch@uriah.heep.sax.de
                                   http://www.sax.de/~joerg/

Never trust an operating system you don't have sources for. ;-)