From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266466AbUGPAPz (ORCPT ); Thu, 15 Jul 2004 20:15:55 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266472AbUGPAPz (ORCPT ); Thu, 15 Jul 2004 20:15:55 -0400 Received: from e3.ny.us.ibm.com ([32.97.182.103]:8590 "EHLO e3.ny.us.ibm.com") by vger.kernel.org with ESMTP id S266466AbUGPAPy (ORCPT ); Thu, 15 Jul 2004 20:15:54 -0400 Date: Thu, 15 Jul 2004 17:14:40 -0700 From: "Martin J. Bligh" To: Jesse Barnes , linux-kernel , Nick Piggin , John Hawkes Subject: Re: [PATCH] reduce inter-node balancing frequency Message-ID: <46970000.1089936880@flay> In-Reply-To: <200407151829.20069.jbarnes@engr.sgi.com> References: <200407151829.20069.jbarnes@engr.sgi.com> X-Mailer: Mulberry/2.1.2 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org > Nick, we've had this patch floating around for awhile now and I'm wondering > what you think. It's needed to boot systems with lots (e.g. 256) nodes, but > could probably be done another way. Do you think we should create a > scheduler domain for every 64 nodes or something? I think that'd make a lot of sense ... > Any other NUMA folks have thoughts about these values? Yeah, change them in arch specific code, not in the global stuff ;-) But seeing as they're dependant (for you) on machine size, as well as arch type, you probably need to do something cleverer in arch_init_sched_domain But the big bugaboo is arch-specific vs general ... we need to break opteron vs i386 vs ia64 out from each other ... they all need different coefficients. If you were going to be really fancy, we could do it in common code off the topology stuff ... but for now, I think it's easier to just set 'em per arch ... M.