From owner-freebsd-smp  Wed Feb  4 08:38:36 1998
Return-Path: <owner-freebsd-smp@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id IAA13208
          for freebsd-smp-outgoing; Wed, 4 Feb 1998 08:38:36 -0800 (PST)
          (envelope-from owner-freebsd-smp@FreeBSD.ORG)
Received: from bsd.synx.com (rt.synx.com [194.167.81.239])
          by hub.freebsd.org (8.8.8/8.8.8) with SMTP id IAA13138
          for <freebsd-smp@freebsd.org>; Wed, 4 Feb 1998 08:38:29 -0800 (PST)
          (envelope-from remy@synx.com)
Received: from s3.synx.com (s3 [192.1.1.247]) by bsd.synx.com (8.6.12/8.6.12) with SMTP id RAA00227 for <freebsd-smp@freebsd.org>; Wed, 4 Feb 1998 17:42:55 +0100
Received: from rs1 by s3.synx.com id aa01426; 4 Feb 98 17:28 GMT
Date: Wed, 4 Feb 1998 17:29:53 -0100 (GMT)
From: Remy NONNENMACHER <remy@synx.com>
To: freebsd-smp@FreeBSD.ORG
Subject: P2L97/DGVS Perfs results and minor problems
Message-ID: <Pine.A32.3.91.980204170648.10787A-100000@rs1>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-freebsd-smp@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org
X-To-Unsubscribe: mail to majordomo@FreeBSD.org "unsubscribe freebsd-smp"


For thoses interested on perfs, I finished setting up an ASUS P2L97-DS, 
128MB SDRAM, 2 x PII 337Mhz (Oc'd 300).

SCSI devs are : 2 x 2940 UW (+ 1 Internal chipset)
Each 2940 chain have 1 x Quantum AtlasII 9Go and 2 x Ibm DGVS 9Go.
(total : 6 x 9Go).

The Ibm's are 10050 rpm, 18MB/s media rate.

Using ccd's, 2 disks give a 33MB/s filesystem (best), and 27MB/s 
(average) reads. Writes seems to be far more slow.

I tested 4 disks ccds and got a 52MB/s peak read. Unfortunetly, this 
raises the 'sys' activity to about 90%.

------ Problems encountered :

- IDE CD attachment. (great problem since it's the only IDE device. The 
ATAPI code do not probe ide0:0, only ide0:1 if no disk found then the 
probing decide a way like : 'no disks ? so no Controller !!)

- RC5 client : (FYI : 930 kk/s/processor). Using the SNAP-971012, I
*NEVER* felt the load of the RC5 clients. Using the 971225, every time the 
RC5 clients are running, all disk activities are slowed (up to 10 time 
slower). Someone heard about a 'nicing' problem ?

------

Congratulations to all the SMP (and other FreeBSD-*) team(s)....