From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S265395AbUIOB3X (ORCPT ); Tue, 14 Sep 2004 21:29:23 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S265795AbUIOB3X (ORCPT ); Tue, 14 Sep 2004 21:29:23 -0400 Received: from mta5.srv.hcvlny.cv.net ([167.206.5.78]:49305 "EHLO mta5.srv.hcvlny.cv.net") by vger.kernel.org with ESMTP id S265395AbUIOB3U (ORCPT ); Tue, 14 Sep 2004 21:29:20 -0400 Date: Tue, 14 Sep 2004 21:29:44 -0400 From: Avi Norowitz Subject: High Memory Results in Poor Performance on Asus P4P800-MX Motherboard To: linux-kernel@vger.kernel.org Message-id: <009f01c49ac3$7b04aa20$0100000a@FIRE> MIME-version: 1.0 X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1441 X-Mailer: Microsoft Outlook Express 6.00.2800.1437 Content-type: text/plain; charset=iso-8859-1 Content-transfer-encoding: 7BIT X-Priority: 3 X-MSMail-priority: Normal Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Greetings, At the dedicated hosting company I work for we have been having some serious performance problems with Linux 2.6.8.1 on Pentium 4 servers running the Asus P4P800-MX motherboard. Some of these servers run as if they were Pentium 1s, while others run as if they were 386s. (I am not exaggerating.) Eventually I found that this problem is resolved with a compiled kernel with high memory disabled. Of course, this is not an attractive solution, since many of our servers have over 1GB of memory. However, I came across this solution on linux-kernel, archived at Google: http://groups.google.com/groups?selm=1bD5H-1aI-9%40gated-at.bofh.it This solution worked for servers with 1GB of memory, but not servers with 1.5GB. However, I was able to modify the commands so it worked with servers with 1.5GB of memory, and presumably servers up to 4GB: echo "disable=9" >| /proc/mtrr echo "disable=8" >| /proc/mtrr echo "disable=7" >| /proc/mtrr echo "disable=6" >| /proc/mtrr echo "disable=5" >| /proc/mtrr echo "disable=4" >| /proc/mtrr echo "disable=3" >| /proc/mtrr echo "disable=2" >| /proc/mtrr echo "disable=1" >| /proc/mtrr echo "disable=0" >| /proc/mtrr echo "base=0x00000000 size=0x100000000 type=write-back" > /proc/mtrr I am very glad that I was able to resolve the problem, but I have three further questions. 1. Is there a better fix to use than to have the above commands run at boot? 2. Which is at fault: the motherboard or the Linux kernel? 3. Should I expect this problem to be resolved in future kernels? If any developers would like root access to a test server having this problem to investigate the problem, please let me know. Thank you for your time. Sincerely, Avi Norowitz Unix Technical Support Fortress ITX / DedicatedNOW / Pegasus Web Technologies