From owner-freebsd-amd64@FreeBSD.ORG Fri Aug 26 12:23:18 2005 Return-Path: X-Original-To: freebsd-amd64@freebsd.org Delivered-To: freebsd-amd64@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8D7B416A41F for ; Fri, 26 Aug 2005 12:23:18 +0000 (GMT) (envelope-from vasim@human-capital.ru) Received: from wolf.resume-bank.ru (mail.resume-bank.ru [62.118.252.60]) by mx1.FreeBSD.org (Postfix) with SMTP id AE96643D46 for ; Fri, 26 Aug 2005 12:23:16 +0000 (GMT) (envelope-from vasim@human-capital.ru) Received: (qmail 74981 invoked by uid 0); 26 Aug 2005 12:23:27 -0000 Received: from unknown (HELO vasimwork) (213.33.147.104) by mail.resume-bank.ru with SMTP; 26 Aug 2005 12:23:27 -0000 Message-ID: <0e5c01c5aa38$a2c4bca0$2107a8c0@vasimwork> From: "Vasim Valejev" To: "Scott Long" References: <00d601c5a88e$1cb8c260$2107a8c0@vasimwork> <4305F56C.8090405@samsco.org> Date: Fri, 26 Aug 2005 16:21:07 +0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1506 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1506 Cc: freebsd-amd64@freebsd.org Subject: Re: Slow PCI-X controller perfomance (adaptec 2130SLP) X-BeenThere: freebsd-amd64@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting FreeBSD to the AMD64 platform List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 26 Aug 2005 12:23:18 -0000 Hi ! > Whenever people complain about slow performance with AAC products and > try to point the finger at the driver, I ask them do repeat their > benchmark under linux and see how it compares. Then they realize just > how good they have it with FreeBSD. I've worked very hard over the > years to make the FreeBSD driver be the fastest driver available for > AAC cards. Yep, you was right, FreeBSD was a lot faster. The linux (Fedora Core, 2.6.11 kernel) got slightly better read speed but writes were 2-3 times slower (about 20-24 MB/s with 64k stripe). Not sure if i did everything correctly under linux but that was really low. > Your particular test is testing the following things: > > 1) disk transfer speed > 2) controller firmware efficiency > 3) cache memory bandwidth on the controller > 4) stripe alignment of the requests > 5) PCI-X bandwidth > > Your disk transfer speed is going to be about 70MB/s per disk. Given > that you are doing RAID-5 on 4 disks, you'll ideally get 3x70MB/s = > 210MB/s. The reason that you multiply by 3 and not 4 is that for a > given stripe, you only have 3 disks that contain data, with the 4th > disk containing parity information. Also note that you are dealing > with Ultra320. Adding more disks than what you have will start to > saturate the bandwidth of the bus (which is probabl between 260-300MB/s > depending on a lot of factors) I've made RAID-5 array (64k stripe) with all six disks but didn't got better than 80..90 MB/s for 2GB file write (on /usr partition with bs=256k) and 133 MB/s read from raw device (first 2GB of /dev/aacd0 with bs=256k, turning off read caching doesn't make any difference). That isn't bad either, so i'll stay with that. But it looks like some kind of bandwidth limit somewhere. Vasim V.