From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Original-To: caml-list@yquem.inria.fr Delivered-To: caml-list@yquem.inria.fr Received: from mail2-relais-roc.national.inria.fr (mail2-relais-roc.national.inria.fr [192.134.164.83]) by yquem.inria.fr (Postfix) with ESMTP id BAA0CBBAF for ; Fri, 4 Dec 2009 20:31:44 +0100 (CET) X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AksBAH7yGEvU4367kWdsb2JhbACcAgEBAQEJCwoHEwO6boQzBI0f X-IronPort-AV: E=Sophos;i="4.47,343,1257116400"; d="scan'208";a="38010709" Received: from moutng.kundenserver.de ([212.227.126.187]) by mail2-smtp-roc.national.inria.fr with ESMTP; 04 Dec 2009 20:31:44 +0100 Received: from st11 (p5B1090BC.dip0.t-ipconnect.de [91.16.144.188]) by mrelayeu.kundenserver.de (node=mreu0) with ESMTP (Nemesis) id 0MHKSt-1NLIRo0xwx-00Du89; Fri, 04 Dec 2009 20:31:43 +0100 Received: from [127.0.0.1] (localhost [127.0.0.1]) by st11 (Postfix) with ESMTP id 66A0B29E64; Fri, 4 Dec 2009 20:31:42 +0100 (CET) Message-ID: <4B19639E.600@hars.de> Date: Fri, 04 Dec 2009 20:31:42 +0100 From: Florian Hars User-Agent: Thunderbird 2.0.0.23 (X11/20090817) MIME-Version: 1.0 To: Aaron Bohannon Cc: caml-list@yquem.inria.fr Subject: Re: [Caml-list] Gc.compact surprisingly helpful References: In-Reply-To: X-Enigmail-Version: 0.95.7 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-Provags-ID: V01U2FsdGVkX1+Jbw/hSG0dZm6a9Za5lW6F2P3Zi2yb9gxu0N0 KH08qDuRKaYxCgW84fUv3lA6c0/NgmLYdljEUp4eV928bnN8qa X7+q0I+ws8= X-Spam: no; 0.00; hars:01 hars:01 compactions:01 -unsafe:01 admirable:01 node:01 resident:98 compile:01 heap:01 caml-list:01 bounds:02 florian:03 florian:03 processors:04 size:95 Aaron Bohannon schrieb: > To my very pleasant surprise, I found the throughput *increased* by about 2%!! > [...] utterly defying my GC intuition. Maybe you stay in L3 cache with the more compact heap, what is your resident set size with and without the additional compactions? Performance on modern processors is a riddle inside an enigma (someone once reported improved performance after removing the -unsafe compile option, which has the single function of making the code faster by removing array bounds checks). - Florian -- But our moral language is fragmented; our contemporaries reject the Kantian hunch that choosing those things most admirable and plausible as ends in themselves is the best practice; autonomous sources of the good are everywhere brown and broken. Thus we have PHP. http://lambda-the-ultimate.org/node/1463