From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 1 Nov 2002 08:32:42 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 1 Nov 2002 08:32:39 -0500 Received: from c17928.thoms1.vic.optusnet.com.au ([210.49.249.29]:51587 "EHLO localhost.localdomain") by vger.kernel.org with ESMTP id convert rfc822-to-8bit; Fri, 1 Nov 2002 08:32:38 -0500 Content-Type: text/plain; charset="us-ascii" From: Con Kolivas To: linux kernel mailing list Subject: autoregulated timeslice added to 2.4.19-ck performance patchset Date: Sat, 2 Nov 2002 00:37:15 +1100 User-Agent: KMail/1.4.3 MIME-Version: 1.0 Content-Transfer-Encoding: 8BIT Message-Id: <200211020037.37457.conman@kolivas.net> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I've added a small patch to the -ck patchset to make it up to ck11 The Hz is set to 1000 (from RML's patch) I've added my own patch which varies the timeslice duration. It is autoregulated by decreasing the duration according to the number of running processes or in the presence of uninterruptible processes. The range of timeslices is now huge depending on demand. The diff against ck10 (a small patch) is available here: http://kernel.kolivas.net Cheers, Con P.S. A brief summary of my patch is here : (from sched.c) MIN_TIMESLICE ( 1 * timeslice_multiplier ) MAX_TIMESLICE ( 30 * timeslice_multiplier ) (from timer.c) timeslice_multiplier autoregulated according to number of running tasks. Starts at 32 and is divided by nr_running()^2. Will drop to 1 if there are any nr_uninterruptible(). SMP is decreased proportionately less (1/smp_num_cpus). -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.0 (GNU/Linux) iD8DBQE9woOPF6dfvkL3i1gRAr61AJ9qW85GQmljDAPKAT4zN9GeeJB3CwCfVxIf Vek+JTftg8FW+hTjkdjR9H0= =DvQf -----END PGP SIGNATURE-----