From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261985AbVANNzq (ORCPT ); Fri, 14 Jan 2005 08:55:46 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261986AbVANNzq (ORCPT ); Fri, 14 Jan 2005 08:55:46 -0500 Received: from gockel.physik3.uni-rostock.de ([139.30.44.16]:3256 "EHLO gockel.physik3.uni-rostock.de") by vger.kernel.org with ESMTP id S261985AbVANNzf (ORCPT ); Fri, 14 Jan 2005 08:55:35 -0500 Date: Fri, 14 Jan 2005 14:55:27 +0100 (CET) From: Tim Schmielau To: Andrew Morton cc: lkml Subject: swapspace layout improvements advocacy In-Reply-To: <20050112105315.2ac21173.akpm@osdl.org> Message-ID: References: <20050112123524.GA12843@lnx-holt.americas.sgi.com> <20050112105315.2ac21173.akpm@osdl.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 12 Jan 2005, Andrew Morton wrote: > Our current way of allocating swap can cause us to end up with little > correlation between adjacent pages on-disk. But this can be improved. THe > old swapspace-layout-improvements patch was designed to fix that up, but > needs more testing and tuning. > > It clusters pages on-disk via their virtual address. 2.6 seems in due need of such a patch. I recently found out that 2.6 kernels degrade horribly when going into swap. On my dual PIII-850 with as little as 256 mb ram, I can easily demonstrate that by opening about 40-50 instances of konquerer with large tables, many images and such things. When the machine is into 80-120 mb of the 256 mb swap partition, it becomes almost unusable. Even the desktop background picture needs ~20sec to update, not to talk about any windows' contents. And you can literally hear the reason for it: the harddisk is seeking like crazy. I've applied Ingo Molnars swapspace-layout-improvements-2.6.9-rc1-bk12-A1 port of the patch to a 2.6.11-rc1 kernel, and it handles the same workload much smoother. It's slow, but you can work with it. I just wonder why noone else complained yet. Are systems with tight memory constraints so uncommon these days? Tim