Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1760830AbXEUI3a (ORCPT ); Mon, 21 May 2007 04:29:30 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1760288AbXEUI3J (ORCPT ); Mon, 21 May 2007 04:29:09 -0400 Received: from holomorphy.com ([66.93.40.71]:46398 "EHLO holomorphy.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757170AbXEUI3I (ORCPT ); Mon, 21 May 2007 04:29:08 -0400 Date: Mon, 21 May 2007 01:29:26 -0700 From: William Lee Irwin III To: Dmitry Adamushko Cc: Peter Williams , Ingo Molnar , Linux Kernel Subject: Re: [patch] CFS scheduler, -v12 Message-ID: <20070521082926.GH19966@holomorphy.com> References: <20070513153853.GA19846@elte.hu> <464A6698.3080400@bigpond.net.au> <20070516063625.GA9058@elte.hu> <464CE8FD.4070205@bigpond.net.au> <20070518071325.GB28702@elte.hu> <464DA61A.4040406@bigpond.net.au> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Organization: The Domain of Holomorphy User-Agent: Mutt/1.5.13 (2006-08-11) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1460 Lines: 30 On Sat, May 19, 2007 at 03:27:54PM +0200, Dmitry Adamushko wrote: > Just an(quick) another idea. Say, the load balancer would consider not > only p->load_weight but also something like Tw(task) = > (time_spent_on_runqueue / total_task's_runtime) * some_scale_constant > as an additional "load" component (OTOH, when a task starts, it takes > some time for this parameter to become meaningful). I guess, it could > address the scenarios your have described (but maybe break some others > as well :) ... > Any hints on why it's stupid? I guess I'll take time out from coding to chime in. cfs should probably consider aggregate lag as opposed to aggregate weighted load. Mainline's convergence to proper CPU bandwidth distributions on SMP (e.g. N+1 tasks of equal nice on N cpus) is incredibly slow and probably also fragile in the presence of arrivals and departures partly because of this. Tong Li's DWRR repairs the deficit in mainline by synchronizing epochs or otherwise bounding epoch dispersion. This doesn't directly translate to cfs. In cfs cpu should probably try to figure out if its aggregate lag (e.g. via minimax) is above or below average, and push to or pull from the other half accordingly. -- wli - 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/