[2512] in linux-scsi channel archive
Re: Adaptec 2940 support
daemon@ATHENA.MIT.EDU (Jeff Noxon)
Mon Sep 22 13:58:48 1997
Date: Mon, 22 Sep 1997 12:13:21 -0500
From: Jeff Noxon <jeff@planetfall.com>
To: Zlatko.Calusic@CARNet.hr
Cc: Jeff Noxon <jeff@planetfall.com>, linux-scsi@vger.rutgers.edu
In-Reply-To: <87raahffzl.fsf@atlas.infra.CARNet.hr>; from Zlatko Calusic on Mon, Sep 22, 1997 at 06:52:30PM +0200
On Mon, Sep 22, 1997 at 06:52:30PM +0200, Zlatko Calusic wrote:
> Jeff Noxon <jeff@planetfall.com> writes:
> > The driver is not good at recovering from error conditions, it has many
> > bugs, and it could probably be faster.
>
> Well, this is what concerns me the most. I hope that I didn't throw
> lots of money on something that will in the end prove as unstable or
> of bad performance. Of course, Linux driver quality is of the biggest
> importance here, and I hope it will not fail in accomplishing it's
> task. :)
Don't enable the /proc stuff, don't enable any extra features you don't
have to (tagged queuing is probably ok though), and you should be fine.
You can get away with more if you install some of the latest patches. I
am running 2.0.31-pre9 without any patches, and it is stable.
> As I can see, too, driver is improving constantly. I'm just hoping it's
> not because it's unstable. :)
Other than the above, I think most of the problems are with unstable bus
conditions, resets, timeouts and the like. I put all my other devices
on a 1542CF in the meantime to avoid having to deal with trouble.
> My IBM-DAQA33240 (3.2GB EIDE) performs at about 4.5 MB/sec, which I
> find fast enough, but I'm not satisfied with CPU consumption, of
> course. I believe that the biggest difference between my Pentium Linux
> (133Mhz) and UltraSparc Solaris (143Mhz) is in quality of I/O
> subsystem (SCSI <-> EIDE).
Have you tried IRQ unmasking? Enabling caches and 32-bit DMA?
> I hope that I'm right, and that connecting SCSI disk will greatly
> improve perception of speed on my (beloved :) ) Linux machine.
Probably, but try tweaking the IDE parameters with hdparm if you haven't
done so already. And keep in mind that your CPU utilization figures are
not terrible considering the data being pushed around.
This is the bonnie result with my raid0 array on twin wide SCSI channels:
-------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
100 4297 95.3 7597 62.9 3045 48.9 4188 85.3 11228 71.4 102.6 7.4
/proc/mdstat:
Personalities : [1 linear] [2 raid0]
read_ahead 120 sectors
md0 : active raid0 sdb1 sdc1 4403168 blocks 4k chunks
md1 : inactive
md2 : inactive
md3 : inactive
The CPU utilization is negligible when you're not banging away on the disks
the way Bonnie does.
Jeff
