From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750929AbWEIAnX (ORCPT ); Mon, 8 May 2006 20:43:23 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750926AbWEIAnX (ORCPT ); Mon, 8 May 2006 20:43:23 -0400 Received: from mail21.syd.optusnet.com.au ([211.29.133.158]:50340 "EHLO mail21.syd.optusnet.com.au") by vger.kernel.org with ESMTP id S1750827AbWEIAnX (ORCPT ); Mon, 8 May 2006 20:43:23 -0400 References: <1147130298.30649.33.camel@localhost.localdomain> Message-ID: X-Mailer: http://www.courier-mta.org/cone/ From: Con Kolivas To: tim.c.chen@linux.intel.com Cc: linux-kernel@vger.kernel.org, mingo@elte.hu Subject: Re: Regression seen for patch =?ISO-8859-1?Q?=22sched:dont?= decrease idle sleep =?ISO-8859-1?B?YXZnIg==?= Date: Tue, 09 May 2006 10:43:09 +1000 Mime-Version: 1.0 Content-Type: multipart/signed; boundary="=_mimegpg-kolivas.org-32203-1147135389-0001"; micalg=pgp-sha1; protocol="application/pgp-signature" Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org This is a MIME GnuPG-signed message. If you see this text, it means that your E-mail or Usenet software does not support MIME signed messages. --=_mimegpg-kolivas.org-32203-1147135389-0001 Content-Type: text/plain; format=flowed; charset="US-ASCII" Content-Disposition: inline Content-Transfer-Encoding: 7bit Tim Chen writes: > Con, > > As a result of the patch "sched:dont decrease idle sleep avg" > introduced after 2.6.15, there is a 4% drop in Volanomark > throughput on our Itanium test machine. > Probably the following happened: > Compared to previous code, this patch slightly increases > the the priority boost when a job is woken up. > This adds priority spread and variations to the wait time of jobs > on run queue if we have a lot of similar jobs in the system. > > See patch: > http://www.kernel.org/git/? > p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=e72ff0bb2c163eb13014ba113701bd42dab382fe Lovely This patch corrects a bug in the original code which unintentionally dropped the priority of tasks that were idle but were already high priority on other merits. It doesn't further increase the priority. The 4% almost certainly is due to the lack of any locking in the threading model used by the java virtual machine on volanomark and it being pure luck that penalising particularly idle tasks previously improved the wakeup timing of basically yielding dependant threads. This patch did fix bugs related to interactive yet idle tasks like consoles misbehaving. The fact that the presence of that particular bug improved a multithreaded benchmark that uses such a threading model is pure chance and (obviously) not design. I wouldn't like to see this bug reintroduced on the basis of this benchmark result. --- -ck --=_mimegpg-kolivas.org-32203-1147135389-0001 Content-Type: application/pgp-signature Content-Transfer-Encoding: 7bit -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQBEX+WdZUg7+tp6mRURAibSAJwLm8VyjxrmG/chjdrwQxXWM9K8kQCdHRcT 6tgKkyVkvF01kucNU9H0z9g= =b+iT -----END PGP SIGNATURE----- --=_mimegpg-kolivas.org-32203-1147135389-0001--