From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753132AbXDNPSA (ORCPT ); Sat, 14 Apr 2007 11:18:00 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753430AbXDNPSA (ORCPT ); Sat, 14 Apr 2007 11:18:00 -0400 Received: from ottawa-hs-64-26-128-89.s-ip.magma.ca ([64.26.128.89]:3598 "EHLO mail.rtr.ca" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753132AbXDNPSA (ORCPT ); Sat, 14 Apr 2007 11:18:00 -0400 Message-ID: <4620F0A5.9060707@rtr.ca> Date: Sat, 14 Apr 2007 11:17:57 -0400 From: Mark Lord User-Agent: Thunderbird 1.5.0.10 (X11/20070221) MIME-Version: 1.0 To: Ingo Molnar Cc: Willy Tarreau , Nick Piggin , linux-kernel@vger.kernel.org, Linus Torvalds , Andrew Morton , Con Kolivas , Mike Galbraith , Arjan van de Ven , Thomas Gleixner Subject: Re: [Announce] [patch] Modular Scheduler Core and Completely Fair Scheduler [CFS] References: <20070413202100.GA9957@elte.hu> <20070414020424.GB14544@wotan.suse.de> <20070414063254.GB14875@elte.hu> <20070414064334.GA19463@elte.hu> <20070414080833.GL943@1wt.eu> <20070414083625.GM943@1wt.eu> <20070414105338.GB19454@elte.hu> In-Reply-To: <20070414105338.GB19454@elte.hu> 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 Ingo Molnar wrote: > i kept the 50%/50% rule from the old scheduler, but maybe it's a more > pristine (and smaller/faster) approach to just not give new children any > stats history to begin with. I've implemented an add-on patch that > implements this, you can find it at: > > http://redhat.com/~mingo/cfs-scheduler/sched-fair-fork.patch I've been running my desktop (single-core Pentium-M w/2GB RAM, Kubuntu Dapper) with the new CFS for much of this morning now, with the odd switch back to the stock scheduler for comparison. Here, CFS really works and feels better than the stock scheduler. Even with a "make -j2" kernel rebuild happening (no manual renice, either!) things "just work" about as smoothly as ever. That's something which RSDL never achieved for me, though I have not retested RSDL beyond v0.34 or so. Well done, Ingo! I *want* this as my default scheduler. Things seemed slightly less smooth when I had the CPU hogs and fair-fork extension patches both applied. I'm going to try again now with just the fair-fork added on. Cheers Mark