From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S269938AbUJNAy5 (ORCPT ); Wed, 13 Oct 2004 20:54:57 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S269939AbUJNAy5 (ORCPT ); Wed, 13 Oct 2004 20:54:57 -0400 Received: from omx3-ext.sgi.com ([192.48.171.20]:7298 "EHLO omx3.sgi.com") by vger.kernel.org with ESMTP id S269938AbUJNAyz convert rfc822-to-8bit (ORCPT ); Wed, 13 Oct 2004 20:54:55 -0400 Date: Thu, 14 Oct 2004 10:53:00 +1000 From: Nathan Scott To: Nick Piggin , Andrew Morton Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org, linux-xfs@oss.sgi.com Subject: Re: Page cache write performance issue Message-ID: <20041014005300.GA716@frodo> References: <20041013054452.GB1618@frodo> <20041012231945.2aff9a00.akpm@osdl.org> <20041013063955.GA2079@frodo> <20041013000206.680132ad.akpm@osdl.org> <20041013172352.B4917536@wobbly.melbourne.sgi.com> <416CE423.3000607@cyberone.com.au> <20041013013941.49693816.akpm@osdl.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: 8BIT In-Reply-To: <20041013013941.49693816.akpm@osdl.org> User-Agent: Mutt/1.5.3i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Oct 13, 2004 at 01:39:41AM -0700, Andrew Morton wrote: > Nick Piggin wrote: > > > > Andrew probably has better ideas. > > uh, is this an ia32 highmem box? Yep, it is. > If so, you've hit the VM sour spot. > ... > Basically, *any* other config is fine. 896MB and below, 1.5GB and above. I just tried switching CONFIG_HIGHMEM off, and so running the machine with 512MB; then adjusted the test to write 256M into the page cache, again in 1K sequential chunks. A similar mis- behaviour happens, though the numbers are slightly better (up from ~4 to ~6.5MB/sec). Both ext2 and xfs see this. When I drop the file size down to 128M with this kernel, I see good results again (as we'd expect). I'm being pulled onto other issues atm, but in the background I could try reverting specific changesets if you guys can suggest anything in particular that might be triggering this? thanks! -- Nathan