From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261914AbTJAFSd (ORCPT ); Wed, 1 Oct 2003 01:18:33 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261928AbTJAFSd (ORCPT ); Wed, 1 Oct 2003 01:18:33 -0400 Received: from stroke.of.genius.brain.org ([206.80.113.1]:41442 "EHLO stroke.of.genius.brain.org") by vger.kernel.org with ESMTP id S261914AbTJAFSc (ORCPT ); Wed, 1 Oct 2003 01:18:32 -0400 Date: Wed, 1 Oct 2003 01:18:27 -0400 From: "Murray J. Root" To: linux-kernel@vger.kernel.org Subject: Re: 2.6.0-test6 scheduling(?) oddness Message-ID: <20031001051827.GE1416@Master> Mail-Followup-To: linux-kernel@vger.kernel.org References: <20031001032238.GB1416@Master> <20030930215512.1df59be3.akpm@osdl.org> <3F7A604F.1060905@cyberone.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3F7A604F.1060905@cyberone.com.au> User-Agent: Mutt/1.4.1i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Oct 01, 2003 at 03:04:15PM +1000, Nick Piggin wrote: > > > Andrew Morton wrote: > > >"Murray J. Root" wrote: > > > >>The render finishes in the same 30 minutes, then oowriter starts. > >>oowriter takes about 3 seconds to load if no rendering is going on. > >> > > > >OpenOffice uses sched_yield() in strange ways which causes it to > >get hopelessly starved on 2.6 kernels. I think RH have a fixed version, > >but I don't know if that has propagated into the upstream yet. > > > > > > OK. Aside from the OpenOffice issue, you still have povray taking 50% > longer to complete, which is quite remarkable if its single threaded and > nothing else is running. Maybe its not the scheduler change? Anyway, > Con will want to know exactly which version of his scheduler you used in > test5 to check for possibilities. > patch-test5-O20int Not sure what caused the increase from 2.6.0-test5 to 2.6.0-test6. One noticeable thing - I get a noticeable speed increase if I turn off rendering to the screen, so that it only goes to a file. I don't get that speed improvement in 2.5.65 - only in 2.6.0-test5 and 2.6.0-test6 test5 goes from 20 mins to about 15 mins (the same speed as 2.5.65 with or without screen rendering) test6 goes from 30 mins to 24 mins - still worse than test5 by a lot. -- Murray J. Root