From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752028AbXBVXIZ (ORCPT ); Thu, 22 Feb 2007 18:08:25 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752034AbXBVXIZ (ORCPT ); Thu, 22 Feb 2007 18:08:25 -0500 Received: from mga07.intel.com ([143.182.124.22]:37169 "EHLO azsmga101.ch.intel.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752028AbXBVXIY (ORCPT ); Thu, 22 Feb 2007 18:08:24 -0500 X-ExtLoop1: 1 X-IronPort-AV: i="4.14,207,1170662400"; d="scan'208"; a="185906957:sNHT50141560" Date: Thu, 22 Feb 2007 14:33:00 -0800 From: "Siddha, Suresh B" To: Nick Piggin Cc: "Siddha, Suresh B" , mingo@elte.hu, linux-kernel@vger.kernel.org, rostedt@goodmis.org Subject: Re: [patch 2/2] sched: dynticks idle load balancing - v2 Message-ID: <20070222143259.B14789@unix-os.sc.intel.com> References: <20061213150314.B12795@unix-os.sc.intel.com> <20061213233157.GA20470@elte.hu> <20061213151926.C12795@unix-os.sc.intel.com> <20061219201247.GA12648@elte.hu> <20061219131223.E23105@unix-os.sc.intel.com> <20070116113505.GA6294@elte.hu> <20070130135709.B32010@unix-os.sc.intel.com> <20070216180335.A8744@unix-os.sc.intel.com> <20070216180842.B8744@unix-os.sc.intel.com> <20070222032654.GD12137@wotan.suse.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <20070222032654.GD12137@wotan.suse.de>; from npiggin@suse.de on Thu, Feb 22, 2007 at 04:26:54AM +0100 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Feb 22, 2007 at 04:26:54AM +0100, Nick Piggin wrote: > This is really ugly, sorry :( hm. myself and others too thought it was a simple and nice idea. > My suggestion for handling this was to increase the maximum balance > interval for an idle CPU, and just implement a global shutdown when > the entire system goes idle. > > The former should take care of the power savings issues for bare metal > hardware, and the latter should solve performance problems for many idle > SMP guests. It should take very little code to implement. coming to max balance interval will be challenging. It needs to save power and at the same time respond to load changes fast enough. > If that approach doesn't cut it, then at least we can have some numbers > to see how much better yours is so we can justify including it. > > If you are against my approach, then I can have a try at coding it up > if you like? Sure. If you can provide a patch, I will be glad to provide power and performance comparision numbers with both the approaches. thanks, suresh