From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751536AbWLOLEW (ORCPT ); Fri, 15 Dec 2006 06:04:22 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751598AbWLOLEW (ORCPT ); Fri, 15 Dec 2006 06:04:22 -0500 Received: from smtp41.singnet.com.sg ([165.21.103.142]:35068 "EHLO smtp41.singnet.com.sg" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751536AbWLOLEV (ORCPT ); Fri, 15 Dec 2006 06:04:21 -0500 Message-ID: <4582815F.6040906@homeurl.co.uk> Date: Fri, 15 Dec 2006 19:05:03 +0800 From: Bob User-Agent: Thunderbird 1.5.0.8 (Windows/20061025) MIME-Version: 1.0 To: linux-kernel@vger.kernel.org Subject: Re: Need to enable caches in SMP ? (was Kernel 2.6 SMP very slow with ServerWorks LE Chipset) References: <4577AA11.6020906@homeurl.co.uk> <4580C054.2080902@homeurl.co.uk> <20061214110324.780b4bf0@localhost.localdomain> <458189FD.2080805@wolfmountaingroup.com> In-Reply-To: <458189FD.2080805@wolfmountaingroup.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Jeff V. Merkey wrote: > Alan wrote: >>> As per Alan's suggestion I decompressed the kernel source tree with >>> the processes pegged to one CPU then the other, and as he predicted >>> it took vastly longer on one CPU than the other, but I don't know >>> what that implies, or how to fix it. >> >>> From the timing it sounds like one processor cache is disabled which >>> is a >> little peculiar to say the least. >> > enable the L1 cache in the processor. BIOS settings, no doubt. > > Jeff The very spartan Phoenix BIOS doesn't have any options to enable or disable CPU Cache, which I know full well to enable, but it does have something rather vaguely called "Memory Caching" the enabling of which seems to have fixed the problem, it's strange it only disabled the cache on one CPU and only under 2.6, I'll investigate more in January. It's good to have the fix search able, this thread would have saved me from making an ass of myself. Thank you for your help.