From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by befuddled.reisers.ca (Postfix, from userid 65534) id 70F851EF701; Tue, 1 Jul 2014 16:51:23 -0400 (EDT) Received: from mta1.math.wisc.edu (mta1.math.wisc.edu [144.92.166.23]) by befuddled.reisers.ca (Postfix) with ESMTP id D04E71EF6F8 for ; Tue, 1 Jul 2014 16:51:21 -0400 (EDT) Received: from localhost (localhost [127.0.0.1]) by mta1.math.wisc.edu (Postfix) with ESMTP id AFD19388776 for ; Tue, 1 Jul 2014 15:51:16 -0500 (CDT) Received: from mta1.math.wisc.edu ([127.0.0.1]) by localhost (mta1.math.wisc.edu [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 4v2cpaTkmd4M for ; Tue, 1 Jul 2014 15:51:16 -0500 (CDT) Received: from mta1.math.wisc.edu (localhost [127.0.0.1]) by mta1.math.wisc.edu (Postfix) with ESMTP id D699238885B for ; Tue, 1 Jul 2014 15:51:15 -0500 (CDT) X-Spam-Checker-Version: SpamAssassin 3.3.2 (2011-06-06) on mta1.math.wisc.edu X-Spam-Level: X-Spam-Status: No, score=-101.3 required=6.5 tests=ALL_TRUSTED,BAYES_05, USER_IN_WHITELIST autolearn=disabled version=3.3.2 Received: from mailhost.math.wisc.edu (erdos.math.wisc.edu [144.92.166.25]) by mta1.math.wisc.edu (Postfix) with ESMTP for ; Tue, 1 Jul 2014 15:51:15 -0500 (CDT) Received: from [144.92.166.19] (vv507j.math.wisc.edu [144.92.166.19]) (using TLSv1 with cipher ECDHE-RSA-AES128-SHA (128/128 bits)) (No client certificate requested) by mailhost.math.wisc.edu (Postfix) with ESMTPSA id D383D4216B9 for ; Tue, 1 Jul 2014 15:51:15 -0500 (CDT) Message-ID: <53B31F42.3060306@math.wisc.edu> Date: Tue, 01 Jul 2014 15:51:14 -0500 From: "John G. Heim" Organization: University of Wisconsin-Madison User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.6.0 MIME-Version: 1.0 To: "Speakup is a screen review system for Linux." Subject: Re: small super computer References: <53AF79A2.5030502@math.wisc.edu> <20140629023754.GA4657@localhost.localdomain> <53B028D2.5030601@math.wisc.edu> <20140629165747.GA3481@localhost.localdomain> <53B17D3B.8010006@math.wisc.edu> <53B2A81E.8030000@math.wisc.edu> <20140701145820.GA4643@localhost.localdomain> <20140701164847.GA3289@localhost.localdomain> In-Reply-To: <20140701164847.GA3289@localhost.localdomain> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 X-BeenThere: speakup@linux-speakup.org X-Mailman-Version: 2.1.16 Precedence: list Reply-To: "Speakup is a screen review system for Linux." List-Id: "Speakup is a screen review system for Linux." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 01 Jul 2014 20:51:23 -0000 Cloud computeing is good for high throughput computing. It is not good for high performance computing. The difference is response time. If you have a process that is going to take eleventy gazillion CPU cycles and you need the answer in an hour, you can use cloud computing. But if you are running aprocess that takes only one gazillion cpu cycles but you need the answer now, that's high performance computing. As I said, I don't believe you need super computing at all for what you want to do but if you did, it'd be high performance computing. Something like a game requires high performance computing. Something like DNA sequencing requires high throughput computing. You thing is more like a game than like gne sequencing. If you use cloud computing, you're going to ask your machine a question and then it would send the question out to the cloud over your slow ethernet connection, and I don't care how fast your network connection is, it's still slow comppared to the speed of the bus inside a computer. Then you have to sit there and wait for the answer. The small super computers we have wher (32 cores) don't have a cluster connection at all. If you want to run your process on more than 32 cores at a time you can't do it. But most super computers are made up f dozens of nodes and they are clustered together with a very, very fast ethernet. You can do high throughput computing on a slow ethernet cluster but you can't do high performance computing. On 07/01/14 11:48, Doug Smith wrote: > > > That's a good idea. I wonder if they are using something similar to what we've been talking about to make up their cloud. > > I might have to look into this in the future and that would be good to have that experience. > > > > Thanks. > > > > -- --- John G. Heim, 608-263-4189, jheim@math.wisc.edu