From mboxrd@z Thu Jan 1 00:00:00 1970 From: Gottfried Haider Subject: Re: sata link speed on amd hudson Date: Wed, 13 Apr 2011 14:32:44 +0200 Message-ID: References: <20110403092911.GA5998@liondog.tnic> <1E8B869C0C6913418421A406C094DF7C011FEC7C@sshaexmb1.amd.com> <1E8B869C0C6913418421A406C094DF7C011FF4A1@sshaexmb1.amd.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Return-path: Received: from mail-iw0-f174.google.com ([209.85.214.174]:37676 "EHLO mail-iw0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750901Ab1DMMdE (ORCPT ); Wed, 13 Apr 2011 08:33:04 -0400 Received: by iwn34 with SMTP id 34so544439iwn.19 for ; Wed, 13 Apr 2011 05:33:04 -0700 (PDT) In-Reply-To: <1E8B869C0C6913418421A406C094DF7C011FF4A1@sshaexmb1.amd.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: "Huang, Shane" Cc: Borislav Petkov , linux-ide@vger.kernel.org, "Petkov, Borislav" Hi Shane, >> Can this bit be ignored/overwritten by the driver in theory? > I'm not quite sure, but anyway driver should not ignore/overwrite > it to modify the platform's policy, which was decided by OEM and > verified. We don't know any regression issue either. I understand the reasoning, just want to make clear that the disk performance I am seeing is way below what I had expected. Now I don't know if there is SATA 3 Gbps to blame or not - but since both the chipset and the drive claim to support it _and_ the drive's specs would warrant 6 Gbps (afaics), I'd really want to give that a try. running hdparm -t /dev/sda on a new Intel SSD 510 (120GB) gives here: Timing buffered disk reads: 360 MB in 3.01 seconds = 119.66 MB/sec The drive has an advertised sequential read performance of "up to 400 MB/s (6Gbps)" and a real-world test (on Anandtech) shows at least 232.1 MB/s of that.. Any suggestions?