From owner-freebsd-hackers@FreeBSD.ORG Thu Jul 31 06:11:50 2003 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 6957537B401 for ; Thu, 31 Jul 2003 06:11:50 -0700 (PDT) Received: from mail1.co.ru (mail1.co.ru [194.85.128.29]) by mx1.FreeBSD.org (Postfix) with ESMTP id 170F543F3F for ; Thu, 31 Jul 2003 06:11:49 -0700 (PDT) (envelope-from freebsd1@centrum.cz) Received: from mailhub.co.ru ([194.85.128.15] verified) by mail1.co.ru (CommuniGate Pro SMTP 4.0.6) with ESMTP id 32433742; Thu, 31 Jul 2003 17:11:47 +0400 Received: from 195.16.34.9 (n3-9.dialup.co.ru [195.16.34.9]) by mailhub.co.ru (8.12.9/8.11.0) with ESMTP id h6VDBiwQ028981; Thu, 31 Jul 2003 17:11:46 +0400 (MSD) Date: Thu, 31 Jul 2003 17:11:47 +0400 From: Buckie X-Mailer: The Bat! (v1.62r) X-Priority: 3 (Normal) Message-ID: <1765441567.20030731171147@centrum.cz> To: freebsd-hackers@freebsd.org In-Reply-To: <200307311219.h6VCJLVG053962@spider.deepcore.dk> References: <200307311219.h6VCJLVG053962@spider.deepcore.dk> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable X-MIME-Autoconverted: from 8bit to quoted-printable by mailhub.co.ru id h6VDBiwQ028981 cc: Soeren Schmidt Subject: Re: Ultra ATA card doesn't seem to provide Ultra speeds. X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: Buckie List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 31 Jul 2003 13:11:50 -0000 Soeren, Heh, QNX in first RTP incarnations only had /dev/zero and I sort of get used to it. But nevermind. I've already tried large bsizes (1m and more) and /dev/null too. But I can't really get past the 16Mb/s barrier... Does machine speed really should affect the hard drive performance so much? I don't think that top showed me the processor usage on dd more than 30%, then again it might be something else... I also had a thought about measuring raw io performance, but the only way I know is dd to zero (null). I'm saying this because in... ugh... Windows I get raw io speeds of up to 60(!) Mb per second (HDTach). And that's fine but Windows doesn't want to mount any volumes off this drive if the speed is set to something more than WDMA2 (even UDMA0). It can only access it sector-by-sector, but without any problems though. Overall, the SI0680 card is a mystery! And SiliconImage is silent as well. Any other ideas as to how to measure the performance or ways to speed it up somehow or test it? Z SS> OH, you need to output to /dev/null NOT /dev/zero :) SS> besides that, all looks OK, but you need a fairly large bs to se impr= ovments SS> on this (relatively slow) machine, try a bs of 1m and see if that SS> helps, if not there is something slowing down your system.. SS> -S=F8ren SS> _______________________________________________ SS> freebsd-hackers@freebsd.org mailing list SS> http://lists.freebsd.org/mailman/listinfo/freebsd-hackers SS> To unsubscribe, send any mail to "freebsd-hackers-unsubscribe@freebsd= .org" --=20 Best regards, Buckie mailto:freebsd1@centrum.cz