From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from lists.gentoo.org (pigeon.gentoo.org [208.92.234.80]) by finch.gentoo.org (Postfix) with ESMTP id 70C0C1381F4 for ; Sun, 12 Aug 2012 18:26:43 +0000 (UTC) Received: from pigeon.gentoo.org (localhost [127.0.0.1]) by pigeon.gentoo.org (Postfix) with SMTP id 14A81E0653; Sun, 12 Aug 2012 18:26:14 +0000 (UTC) Received: from smtp.cs.nyu.edu (SMTP.CS.NYU.EDU [128.122.49.97]) by pigeon.gentoo.org (Postfix) with ESMTP id A8079E06B1 for ; Sun, 12 Aug 2012 18:23:15 +0000 (UTC) Received: from ajglap.localdomain (ool-182de1a5.dyn.optonline.net [24.45.225.165]) (authenticated bits=0) by smtp.cs.nyu.edu (8.14.3/8.14.4) with ESMTP id q7CINEj3024262 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT) for ; Sun, 12 Aug 2012 14:23:15 -0400 (EDT) Received: by ajglap.localdomain (Postfix, from userid 1502) id 8E0D770076; Sun, 12 Aug 2012 14:20:58 -0400 (EDT) From: Allan Gottlieb To: gentoo-user@lists.gentoo.org Subject: Re: [gentoo-user] new installation (ssd, new udev, grub2) References: <20120810212213.0ce6e810@khamul.example.com> Date: Sun, 12 Aug 2012 14:20:58 -0400 In-Reply-To: (Paul Hartman's message of "Fri, 10 Aug 2012 15:04:10 -0500") Message-ID: User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/24.1 (gnu/linux) Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-Id: Gentoo Linux mail X-BeenThere: gentoo-user@lists.gentoo.org Reply-to: gentoo-user@lists.gentoo.org MIME-Version: 1.0 Content-Type: text/plain X-Archives-Salt: 4a903889-0e82-4868-925d-6f112891e9e8 X-Archives-Hash: ff170ce343f8d36632a3ad56aa1923d3 On Fri, Aug 10 2012, Paul Hartman wrote: > On Fri, Aug 10, 2012 at 2:22 PM, Alan McKinnon wrote: >> You also don't need an IO scheduler - ssd access is random like >> RAM, no heads moving in and out so no sector ordering to worry about. >> Configure the scheduler as NOOP in kernel config if all drives are ssd's > > I've read some contradictory reports about the best scheduler for SSD. > Apparently some SSD controllers are tuned for sequential reads/writes > (drives that lean heavily on compression, I would imagine?) and for > those drives cfq or deadline may be the better choice. I would try > them all and use whatever works best for your workload. I see. Thanks. allan