Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262770AbUCOXBY (ORCPT ); Mon, 15 Mar 2004 18:01:24 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262667AbUCOXBX (ORCPT ); Mon, 15 Mar 2004 18:01:23 -0500 Received: from phoenix.infradead.org ([213.86.99.234]:25868 "EHLO phoenix.infradead.org") by vger.kernel.org with ESMTP id S262770AbUCOW7k (ORCPT ); Mon, 15 Mar 2004 17:59:40 -0500 Date: Mon, 15 Mar 2004 22:59:39 +0000 From: Christoph Hellwig To: Kurt Garloff , Linux kernel list Subject: Re: dynamic sched timeslices Message-ID: <20040315225939.A23686@infradead.org> Mail-Followup-To: Christoph Hellwig , Kurt Garloff , Linux kernel list References: <20040315224201.GX4452@tpkurt.garloff.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <20040315224201.GX4452@tpkurt.garloff.de>; from garloff@suse.de on Mon, Mar 15, 2004 at 11:42:01PM +0100 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1523 Lines: 35 On Mon, Mar 15, 2004 at 11:42:01PM +0100, Kurt Garloff wrote: > Hi, > > attached patch allows userspace to tune the scheduling timeslices. > It can be used for a couple of things: > * Tune a workload for batch processing: > You'd probably wnat to use long timeslices in order to not reschedule > as often to make good use of your CPU caches > * Tune a workload for interactive use: > Under load, you may want to reduce the scedulilng latencies by using > shorter timeslices (and there are situations where the interactiviy > tweak -- even if they were perfect -- can't save you). > * Tune the ration betweeen maximum and minimum timeslices to make > nice much nicer e.g. > > The patch exports /proc/sys/kernel/max_timeslice and min_timeslice, > unites are us. It also exports HZ (readonly). > The patch implementes the desktop boot parameter which introduces > shorter timeslices. > > Patch is from andrea and is in our 2.4 tree; 2.6 port was done by me and > straightforward. Remove the silly desktop boot parameter and the patch looks basically okay to me. I remember we had a more complete patch to allow tuning the scheduler through sysctls in -mm once, though. Questions is why that one wasn't merged and if the same reasons apply to a 'light' version. - 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/