From owner-freebsd-cluster@FreeBSD.ORG Sun Apr 4 16:30:01 2004 Return-Path: Delivered-To: freebsd-cluster@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0B8A516A4CE for ; Sun, 4 Apr 2004 16:30:01 -0700 (PDT) Received: from catapult.dreamscape.com (catapult.dreamscape.com [206.64.128.85]) by mx1.FreeBSD.org (Postfix) with ESMTP id AB2CD43D2D for ; Sun, 4 Apr 2004 16:30:00 -0700 (PDT) (envelope-from m.mcdonald@computer.org) Received: from mail2.dreamscape.com (mail2.dreamscape.com [206.64.128.18]) i34NTxfJ010779 for ; Sun, 4 Apr 2004 19:30:00 -0400 (EDT) Received: from MICHAELIWZHLNY (pool-141-149-206-40.syr.east.verizon.net [141.149.206.40])i34NTwwS009745 for ; Sun, 4 Apr 2004 19:29:59 -0400 (EDT) Message-ID: <002c01c41aa5$1fbd6b50$2f01a8c0@MICHAELIWZHLNY> From: "Michael McDonald" To: References: <002401c419e7$76692ac0$2f01a8c0@MICHAELIWZHLNY> <20040404113055.GA2677@ipx20050.ipxserver.de> Date: Sun, 4 Apr 2004 20:29:00 -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 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Subject: Re: Request for Cluster Recommendations X-BeenThere: freebsd-cluster@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: Michael McDonald List-Id: Clustering FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 04 Apr 2004 23:30:01 -0000 Thanks for the reponses. I should have made it clearer what my applications will be but you all doped it out pretty well. I'm setting it up for applications in computational chemistry and physics. A lot of eigenvalue and fourier computations. The codes use the netlib LAPACK and blas libraries and FFTW comes up a lot. For the most part, my codes have been data parallel and have invlolved broadcasting parameters and merging results at the end of distributed serial computations. They have involved many evaluations of small matrices and data sets and I get better speedups with simultaneous serial executions. Much of the work I'm looking towards would make use of a grid approach in the style of seti@home or the factoring projects. I would expect bursts of communication separated by periods of computation; Overall, communication wouldn't be so much of a bottleneck, but I'd like it to be fast when it does occur. If Andy Sporner's 30% figure holds up, I think local disks as buffers would allow the network access to be smeared out. I've got 40% in mind as an upper performance limit for ethernet due to collisions, but I can't back that up. Local buffers seem to allow for scheduling comm. so as to avoid collisions. Starting out with 100Mbit may make sense as a cheaper/simpler startup - no NIC purchases. After running some simulation and benchmarking with the apps, upgrading to Gigabit wouldn't be an undue burden. Some initial attempts at estimating communication & compute demands would be in order. Aside from the fiber options, is the cabling the same for 100Mbit and Gbit? Thanks to all, Mike McDonald