From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on yquem.inria.fr X-Spam-Level: X-Spam-Status: No, score=0.0 required=5.0 tests=none autolearn=disabled version=3.1.3 X-Original-To: caml-list@yquem.inria.fr Delivered-To: caml-list@yquem.inria.fr Received: from discorde.inria.fr (discorde.inria.fr [192.93.2.38]) by yquem.inria.fr (Postfix) with ESMTP id 2C4E3BC68 for ; Fri, 27 Oct 2006 23:16:58 +0200 (CEST) Received: from deckard.concept-micro.com (free.concept-micro.com [82.228.75.25]) by discorde.inria.fr (8.13.6/8.13.6) with ESMTP id k9RLGvIv019323 for ; Fri, 27 Oct 2006 23:16:57 +0200 X-Envelope-From: petchema@concept-micro.com X-Envelope-To: Received: from polo.concept-micro.com (root@localhost [127.0.0.1]) by deckard.concept-micro.com (8.13.8/8.13.8/Pierre Etchemaite - 20/05/03) with SMTP id k9RLGkDD027900 for ; Fri, 27 Oct 2006 23:16:47 +0200 Date: Fri, 27 Oct 2006 23:16:45 +0200 From: Pierre =?UTF-8?B?RXRjaGVtYcOvdMOp?= To: caml-list@inria.fr Subject: Re: [Caml-list] caml_oldify_local_roots takes 50% of the total runtime Message-Id: <20061027231645.b68969d2.petchema@concept-micro.com> In-Reply-To: <454206CF.1080306@janestcapital.com> References: <17727.34685.561877.977822@tandem.cs.ru.nl> <1161802911.12050.5.camel@rosella.wigram> <1161870528.20369.27.camel@localhost.localdomain> <454206CF.1080306@janestcapital.com> Organization: Concept Micro 33 SARL X-Mailer: Sylpheed version 2.2.6 (GTK+ 2.6.4; i386-pc-linux-gnu) X-Face: #eTSL0BRng*(!i1R^[)oey6`SJHR{3Sf4dc;"=af8%%;d"%\#"Hh0#lYfJBcm28zu3r^/H ^ d6!9/eElH'p0'*,L3jz_UHGw"+[c1~ceJxAr(^+{(}|DTZ"],r[jSnwQz$/K&@MT^?J#p"n[J>^O[ \ "%*lo](u?0p=T:P9g(ta[hH@uvv X-TRP: honey@concept-micro.com Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-Miltered: at discorde with ID 45427749.001 by Joe's j-chkmail (http://j-chkmail.ensmp.fr)! X-Spam: no; 0.00; oldify:01 runtime:01 citeseer:01 non-issue:01 psu:98 goncalves:98 heap:01 minor:01 minor:01 caml-list:01 pierre:02 caml:02 roots:02 seems:03 seems:03 Le Fri, 27 Oct 2006 09:17:03 -0400, Brian Hurt a écrit : > With respect to 4146 (minor heap size adjusts to memory size)- I'm not > sure this is a good idea. 32K is small enough to fit into L1 cache with > space left over on pretty much all systems these days (64K L1 cache > seems to be standard). Default size is 32K *words*, hence 128kB for 32bit archs and 256kB on 64bits. Small enough to stay in L2 cache, but not L1. Also, from reading http://citeseer.ist.psu.edu/goncalves94cache.html It seems that write-validate cache policy makes minor heaps larger than L2 cache a non-issue for most programs. That paper is 10 years old, did that policy become the norm since then ?