From owner-freebsd-current@FreeBSD.ORG Sun Sep 26 10:24:12 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id B279E16A4CE for ; Sun, 26 Sep 2004 10:24:12 +0000 (GMT) Received: from moutng.kundenserver.de (moutng.kundenserver.de [212.227.126.183]) by mx1.FreeBSD.org (Postfix) with ESMTP id 4724B43D45 for ; Sun, 26 Sep 2004 10:24:12 +0000 (GMT) (envelope-from se@freebsd.org) Received: from [212.227.126.160] (helo=mrelayng.kundenserver.de) by moutng.kundenserver.de with esmtp (Exim 3.35 #1) id 1CBWCR-0007bG-00; Sun, 26 Sep 2004 12:24:11 +0200 Received: from [80.132.243.59] (helo=Gatekeeper.FreeBSD.org) by mrelayng.kundenserver.de with asmtp (Exim 3.35 #1) id 1CBWCQ-0001CG-00; Sun, 26 Sep 2004 12:24:10 +0200 Received: from StefanEsser.FreeBSD.org (StefanEsser [192.168.0.10]) by Gatekeeper.FreeBSD.org (Postfix) with ESMTP id CB3245F19; Sun, 26 Sep 2004 12:24:07 +0200 (CEST) Received: by StefanEsser.FreeBSD.org (Postfix, from userid 200) id 0B7DC2307; Sun, 26 Sep 2004 12:24:07 +0200 (CEST) Date: Sun, 26 Sep 2004 12:24:06 +0200 From: Stefan =?iso-8859-1?Q?E=DFer?= To: Mauro Triulzi Message-ID: <20040926102406.GA2428@StefanEsser.FreeBSD.org> Mail-Followup-To: Stefan =?iso-8859-1?Q?E=DFer?= , Mauro Triulzi , Jeremy Chadwick , freebsd-current@freebsd.org References: <4155DC3F.5070904@freesurf.ch> <20040925214515.GA56471@parodius.com> <41560C61.1040009@freesurf.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <41560C61.1040009@freesurf.ch> User-Agent: Mutt/1.5.6i X-Provags-ID: kundenserver.de abuse@kundenserver.de auth:fa3fae9b6ca38d745862a668565919f6 cc: freebsd-current@freebsd.org cc: Jeremy Chadwick Subject: Re: ATA disk performance (ICH2 controller), some tests andcomparison with Linux 2.6.5 X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 26 Sep 2004 10:24:12 -0000 On 2004-09-26 02:25 +0200, Mauro Triulzi wrote: > -------Sequential Output-------- ---Sequential Input-- --Random-- > -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks--- > Machi MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU % FBSDa 1500 14533 12.8 14184 4.6 9360 3.2 27774 24.0 29067 5.7 119.1 0.6 % FBSDb 1500 27787 24.6 24690 8.3 12519 4.3 24487 21.4 24870 5.0 128.1 0.6 % Linux 1500 26859 97.7 59118 14.0 22480 8.1 22508 78.9 48501 9.0 177.1 0.5 [Table created from results on FreeBSD, FreeBSD after ATA reinit, and Linux.] The difference between the test before and after the reinit appears relevant and very disturbing. May be drive parameters (write cache enable?) are changed be the reinit command, leading to significantly different write performance? Different Bonnie versions use slightly different algortithms. You may want to use the same version (perhaps even the same binary, since the Linux version ought to run under emulation in FreeBSD). Linux used to cache much higher numbers of dirty buffers than FreeBSD, which tried to keep that number low for a number of reasons (e.g. to prevent to high a backlog of writes, but also to avoid active program pages being flushed just to allow for dirty buffers). You may want to test the raw throughput of your drive over the file systems used for the tests. E.g. "dd if=/dev/DISK of=/dev/null bs=64k" with DISK set to the FreeBSD device names of the file systems, where your bonnie test file was placed in each case. Running the same command under both FreeBSD and Linux should give identical results, e.g. the 128MB root partition on my (not really idle) system: # dd if=/dev/ad0s1a bs=64k of=/dev/null 2048+0 records in 2048+0 records out 134217728 bytes transferred in 2.223648 secs (60359249 bytes/sec) Since modern file systems spread data all over the available partition, you have to read the whole partition to get an average raw transfer speed. Be aware, that file system tests (i.e. Bonnie) are highly dependent on the amount of free space in the file system being tested, so if your FreeBSD and Linux partitions had different sizes and amounts of free space, your values aren't really significant ... Regards, STefan