[4747] in linux-scsi channel archive
Re: Really Ssslllllooooooowwwwwww lilo load of kernel. (fwd)
daemon@ATHENA.MIT.EDU (Karl-Heinz Herrmann)
Mon Sep 28 05:48:47 1998
In-Reply-To: <Pine.LNX.3.96.980927212339.348B-100000@filesrv1.baby-dragons.com>
Date: Mon, 28 Sep 1998 11:46:18 +0200 (MET DST)
Reply-To: Karl-Heinz Herrmann <k.-h.herrmann@fz-juelich.de>
From: Karl-Heinz Herrmann <k.-h.herrmann@fz-juelich.de>
To: "Mr. James W. Laferriere" <babydr@baby-dragons.com>
Cc: Linux Scsi maillist <linux-scsi@vger.rutgers.edu>
Hallo!
On 28-Sep-98 Mr. James W. Laferriere wrote:
> Using 'hdparm -tT /dev/sda1'
> the Wide drive gets 'buffered disk reads' from 6.71 to
> 6.75 MB/s, the narrow drive sdb1 gets 3.61 to 3.66
> MB/s
> The above isn't a real test such as bonnie . these
> still
> seem a bit slow for a drive/controller comb. capable of
> bursts to 20 MB/s .
My bonnie Tests suggest that you have quite normal transfer
rates. My 5400rpm drives (one Ulta, one wide) deliver a sustained
rate of 3.5 to 3.8 MB/s on a Adaptec 2940UW adapter.
The raid0 running on them get's indeed allmost the double rate
and the adapter is most probably still not near it's limit. But
with 5400 rpm drives I/you won't get more then maybe 4 MB/s.
The one drive of yours with 6MB should be something faster.
There is a file (readme maybe) coming with bonnie which states
some values of different architectures.
I don't know how good hdparm results are comparable with bonnie
results.
If you can just do reads because of your free disk space, maybe:
date; dd if=/dev/sda of=/dev/null ;date
and a little calculation gives you an idea of sustained read
speed.
Greetings,
K.-H.
----------------------------------------------------------
E-Mail: Karl-Heinz Herrmann <k.-h.herrmann@fz-juelich.de>
Date: 28-Sep-98 Time: 11:39:09
----------------------------------------------------------
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
