Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 19 Mar 2002 09:20:25 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 19 Mar 2002 09:20:15 -0500 Received: from atlas.inria.fr ([138.96.66.22]:23761 "EHLO atlas.inria.fr") by vger.kernel.org with ESMTP id ; Tue, 19 Mar 2002 09:20:00 -0500 Message-Id: <200203191419.g2JEJfM07465@atlas.inria.fr> Content-Type: text/plain; charset="iso-8859-1" From: Nicolas Turro To: Alan Cox , Nicolas.Turro@sophia.inria.fr (Nicolas Turro) Subject: Re: amd nvidia and mem=nopentium Date: Tue, 19 Mar 2002 15:19:40 +0100 X-Mailer: KMail [version 1.3] Cc: linux-kernel@vger.kernel.org In-Reply-To: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Thanks for your answer, Alan. Le Vendredi 15 Mars 2002 20:05, Alan Cox a ?crit : > > Hi, i have system dual athlon XP 1900+ system and a nvidia graphic board > > : I need to use the mem=nopentium kernel parameter in order to run X > > without crashes. I'd like to know : > > 1- what are the consequences of 'mem=nopentium' ? Any performance loss ? > > Yes. On the whole probably not a lot. You are running XP not MP processors > and the like so you are obviously not too worried about stability. You > might want to see if it actually does crash without nopentium. Well... In fact, i am really worried about stability, but i don't have much choice on the configuration. I have to choose between this config and a dual 2.0 Ghz Xeon (RDRAM) which has roughtly the same perfs, but which is 50% more expensive ! Do you have pointers showing stability problems when using XP processor in a multiprocessor context ? The Athlon actually crashes as soon as I start X with the nvidia board if i don't use the mem=nopentium option. With a Matrox G450, i don't need this option... > > i intend to use a gigabit ethernet adapter on this box. > > 2- is there any fix going on that i should monitor ? > > Some gige cards don't seem to work with some dual athlon bioses. Other than > that it should be fine I've juste tested an Intel e1000 on it and achieved 900 Mbits/s with ttcp.... So i guess it works... I do some nfs benches right now to see if i can reach the disk transfer limit (around 40 Mo/s for sequencial acces on a big file). N. Turro - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/