From owner-freebsd-hackers Tue Mar 25 10:18:42 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id KAA28698 for hackers-outgoing; Tue, 25 Mar 1997 10:18:42 -0800 (PST) Received: from labinfo.iet.unipi.it (labinfo.iet.unipi.it [131.114.9.5]) by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id KAA28690 for ; Tue, 25 Mar 1997 10:18:35 -0800 (PST) Received: from localhost (luigi@localhost) by labinfo.iet.unipi.it (8.6.5/8.6.5) id SAA05269; Tue, 25 Mar 1997 18:24:03 +0100 From: Luigi Rizzo Message-Id: <199703251724.SAA05269@labinfo.iet.unipi.it> Subject: Re: Compaq Smart-2/P RAID controller results To: md@dcs.qmw.ac.uk (Mark Dawson) Date: Tue, 25 Mar 1997 18:24:03 +0100 (MET) Cc: freebsd-hackers@FreeBSD.org, programmers@dcs.qmw.ac.uk In-Reply-To: from "Mark Dawson" at Mar 25, 97 05:21:27 pm X-Mailer: ELM [version 2.4 PL23] Content-Type: text Sender: owner-hackers@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk > You're right, the Wren-VII disks aren't youthful so the comparison is > not fair to the CCD driver. I've attached the H RAID0 disks to the and not only that, you also had very little memory and probably a not-too-fast controller. > ProLiant's integrated NCR 53c875 wide scsi controller and configured > them as a 6 x 4.3GB CCD disk with the following results: > > -------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 > H CCD 300 8491 76.5 8474 19.1 5123 13.7 10273 96.6 12138 16.2 136.8 2.4 > H RAID0 300 11615 98.3 15231 36.9 9096 27.1 10532 98.0 16073 30.7 131.6 2.6 these are quite interesting results! May I ask you (if it is not too hard) to make a further try with CCD and 2..5 disks, just to see how it scales ? It's hard to have a chance to make measurements in a testbed like yours (with empty disks I guess) ... Another thing one might wonder is why the per-char/per-block performance with CCD are so similar. Perhaps CCD as configured does not exploit enough the available parallelism, or is that limited by the SCSI bandwidth on the NCR53c875 controller ? Thanks Luigi -----------------------------+-------------------------------------- Luigi Rizzo | Dip. di Ingegneria dell'Informazione email: luigi@iet.unipi.it | Universita' di Pisa tel: +39-50-568533 | via Diotisalvi 2, 56126 PISA (Italy) fax: +39-50-568522 | http://www.iet.unipi.it/~luigi/ _____________________________|______________________________________