Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1761058AbXFRQy2 (ORCPT ); Mon, 18 Jun 2007 12:54:28 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752866AbXFRQyV (ORCPT ); Mon, 18 Jun 2007 12:54:21 -0400 Received: from netops-testserver-4-out.sgi.com ([192.48.171.29]:49195 "EHLO relay.sgi.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752539AbXFRQyU (ORCPT ); Mon, 18 Jun 2007 12:54:20 -0400 Date: Mon, 18 Jun 2007 09:54:18 -0700 (PDT) From: Christoph Lameter X-X-Sender: clameter@schroedinger.engr.sgi.com To: Srivatsa Vaddagiri cc: "Paul E. McKenney" , Ingo Molnar , Thomas Gleixner , Dinakar Guniguntala , Dmitry Adamushko , suresh.b.siddha@intel.com, pwil3058@bigpond.net.au, linux-kernel@vger.kernel.org, akpm@linux-foundation.org Subject: Re: v2.6.21.4-rt11 In-Reply-To: <20070618151215.GA9750@linux.vnet.ibm.com> Message-ID: References: <20070613180451.GA16628@elte.hu> <20070613184741.GC8125@linux.vnet.ibm.com> <20070613185522.GA27335@elte.hu> <20070613233910.GJ8125@linux.vnet.ibm.com> <20070615144535.GA12078@elte.hu> <20070615151452.GC9301@linux.vnet.ibm.com> <20070615195545.GA28872@elte.hu> <20070616011605.GH9301@linux.vnet.ibm.com> <20070616084434.GG2559@linux.vnet.ibm.com> <20070616161213.GA2994@linux.vnet.ibm.com> <20070618151215.GA9750@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1166 Lines: 22 On Mon, 18 Jun 2007, Srivatsa Vaddagiri wrote: > This particular machine, elm3b6, is a 4-cpu, (gasp, yes!) 4-node box i.e > each CPU is a node by itself. If you don't have CONFIG_NUMA enabled, > then we won't have cross-node (i.e cross-cpu) load balancing. > Fortunately in your case you had CONFIG_NUMA enabled, but still were > hitting the (gross) load imbalance. > > The problem seems to be with idle_balance(). This particular routine, > invoked by schedule() on a idle cpu, walks up sched-domain hierarchy and > tries to balance in each domain that has SD_BALANCE_NEWIDLE flag set. > The nodes-level domain (SD_NODE_INIT) however doesn't set this flag, > which means idle cpu looks for (im)balance within its own node at most and The nodes-level domain looks for internode balances between up to 16 nodes. It is not restricted to a single node. The balancing on the phys_domain level does balance within a node. - 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/