Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753945AbWKHDEu (ORCPT ); Tue, 7 Nov 2006 22:04:50 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753946AbWKHDEu (ORCPT ); Tue, 7 Nov 2006 22:04:50 -0500 Received: from omx2-ext.sgi.com ([192.48.171.19]:65164 "EHLO omx2.sgi.com") by vger.kernel.org with ESMTP id S1753945AbWKHDEu (ORCPT ); Tue, 7 Nov 2006 22:04:50 -0500 Date: Wed, 8 Nov 2006 14:04:29 +1100 From: David Chinner To: Mikulas Patocka Cc: Albert Cahalan , linux-kernel@vger.kernel.org Subject: Re: 2048 CPUs [was: Re: New filesystem for Linux] Message-ID: <20061108030429.GQ11034@melbourne.sgi.com> References: <787b0d920611041154l69db46abv4c8c467809ada57c@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.2.1i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 994 Lines: 32 On Sat, Nov 04, 2006 at 11:36:49PM +0100, Mikulas Patocka wrote: > >>Does one Linux kernel run on system with 1024 cpus? I guess it > >>must fry spinlocks... (or even lockup due to spinlock livelocks) > > > >The SGI Altix can have 2048 CPUs. > > And does it run one image of Linux? Or more images each on few cpus? One image. > How do they solve problem with spinlock livelocks? By replacing contended spinlocks withsleeping locks, using no-lock techniques (e.g. per-cpu) or changing the algorithm to remove the contention point. w.r.t filesystem locking scalability, you should read this paper: http://oss.sgi.com/projects/xfs/papers/ols2006/ols-2006-paper.pdf Cheers, Dave. -- Dave Chinner Principal Engineer SGI Australian Software Group - 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/