From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail1-relais-roc.national.inria.fr (mail1-relais-roc.national.inria.fr [192.134.164.82]) by yquem.inria.fr (Postfix) with ESMTP id 53C3DBBAF for ; Sun, 18 Jul 2010 00:01:13 +0200 (CEST) X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AiMCAILEQUzRVdY0kGdsb2JhbACfGVEIFQEBAQEJCQwHEQMfrk2CEIR5LohUAQEDBYUgBIN7hFw X-IronPort-AV: E=Sophos;i="4.55,220,1278280800"; d="scan'208";a="63968826" Received: from mail-bw0-f52.google.com ([209.85.214.52]) by mail1-smtp-roc.national.inria.fr with ESMTP; 18 Jul 2010 00:01:12 +0200 Received: by bwz14 with SMTP id 14so2217513bwz.39 for ; Sat, 17 Jul 2010 15:01:12 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:references:in-reply-to :mime-version:content-transfer-encoding:content-type:message-id:cc :x-mailer:from:subject:date:to; bh=xZOY/Q+fY+YZeW0UXQzpYwPm85JDNuAkJ5aVsbc27ek=; b=EDRfTKuUwYhqVdHtIROtygLCwmzccUiz7A94v4rs/0/9/vcWFB/ycQ76pS/2Fb1lMK ORoNMZvHnvAjXCApQGBnkM8K5nV45kBWbkUP7q4HnA5k5TCsucHGSbNpWxq04gFFLrc8 5i9o5REkA/cMhnqoTq6lcJdGLsZp/8TBJEDvQ= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=references:in-reply-to:mime-version:content-transfer-encoding :content-type:message-id:cc:x-mailer:from:subject:date:to; b=WJx7U9ZVPdI3wRH9K6kNtU889zfkqcGAT0Gouh6dp1q22tIL0BOvYSKCyysHB/S5wS OOwaHObfkQGEun4lUucva3QugS8IjBzq2vr3KgQ2XdXFjItPlTjl8n/KVIwHuk1K0hhE iMkbF7sMUwPipi6HGQijAwzSUwK9ZpYyfyCmc= Received: by 10.204.118.3 with SMTP id t3mr2298766bkq.163.1279404071988; Sat, 17 Jul 2010 15:01:11 -0700 (PDT) Received: from [77.67.215.126] ([77.67.215.126]) by mx.google.com with ESMTPS id g11sm17648285bkw.22.2010.07.17.15.01.08 (version=TLSv1/SSLv3 cipher=RC4-MD5); Sat, 17 Jul 2010 15:01:11 -0700 (PDT) References: <87sk3mcaeq.fsf@frosties.localdomain> <87zkxsfvsg.fsf@frosties.localdomain> <87hbk0vzut.fsf@frosties.localdomain> <942FD9F6-601C-4F02-83E8-84CD280A04FF@gmail.com> <87tynxc35v.fsf@frosties.localdomain> In-Reply-To: <87tynxc35v.fsf@frosties.localdomain> Mime-Version: 1.0 (iPhone Mail 8A293) Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=us-ascii Message-Id: Cc: Goswin von Brederlow , Rich Neswold , "caml-list@yquem.inria.fr" X-Mailer: iPhone Mail (8A293) From: Eray Ozkural Subject: Re: [Caml-list] Smart ways to implement worker threads Date: Sun, 18 Jul 2010 01:00:50 +0300 To: Goswin von Brederlow X-Spam: no; 0.00; eray:01 ozkural:01 speedup:01 algos:01 eray:01 threads:01 caml-list:01 data:02 data:02 algorithms:03 seems:03 apps:04 apps:04 optimized:04 nvidia:94 It looks like the multi core architectures will be around for a while only t= o be superseded by more advanced parallel micro-architectures. Of course computation-bound apps will benefit more but this does not mean th= at memory intensive apps will not run well on multi core architectures. Care= fully optimized code will run great on multi core architectures with high me= mory bandwidth. Depending on the application of course. Right now it seems to me that shared mem is more suitable for data mining ap= ps (and many other including scientific computing apps). Data mining has nev= er been embarrassingly parallel and it's difficult to get good speedup on di= stributed architectures. I should know I've worked on the parallel solution o= f two such problems. Provided that the space complexity is manageable, I thi= nk such network-bound cluster algorithms have more efficient counter-parts o= n multi-core. In data mining algos some data may have to be replicated or co= mmunicated. If the space doesn't blow up, NVIDIA Fermi would be ideal for th= ose problems. No more copying. On the new Tesla max memory will be 6x4=3D24g= b. For all I care Intel is obsolete until they give me a thousand cores. Shared memory support is essential at any rate. I don't see why we can't hav= e it in an excellent way! Best, Eray