Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752293AbWAFSFl (ORCPT ); Fri, 6 Jan 2006 13:05:41 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752155AbWAFSFk (ORCPT ); Fri, 6 Jan 2006 13:05:40 -0500 Received: from omx3-ext.sgi.com ([192.48.171.26]:23763 "EHLO omx3.sgi.com") by vger.kernel.org with ESMTP id S1751413AbWAFSFk (ORCPT ); Fri, 6 Jan 2006 13:05:40 -0500 Date: Fri, 6 Jan 2006 10:04:56 -0800 (PST) From: Christoph Lameter To: Matthew Wilcox cc: "Luck, Tony" , Arjan van de Ven , hawkes@sgi.com, Tony Luck , Andrew Morton , linux-ia64@vger.kernel.org, linux-kernel@vger.kernel.org, Jack Steiner , Dan Higgins , John Hesterberg , Greg Edwards Subject: Re: [PATCH] ia64: change defconfig to NR_CPUS==1024 In-Reply-To: <20060106174957.GF19769@parisc-linux.org> Message-ID: References: <20060106174957.GF19769@parisc-linux.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 Content-Length: 2262 Lines: 43 On Fri, 6 Jan 2006, Matthew Wilcox wrote: > On Fri, Jan 06, 2006 at 09:45:20AM -0800, Luck, Tony wrote: > > I suppose that depends on your expectations from defconfig. In my > > mind its the one that builds into a kernel that will boot and run > > on just about any box. People who want to get a bit of extra performance > > will do the re-compilation to strip out the bits that they don't want > > and tune down limits that are set higher than they need. I only > > You can use that argument to set the CPU limit low too -- since a kernel > with a CPU limit lower than the number of CPUs in the box will just ignore > the additional ones, people who want to get the additional performance > will tune limits that are set lower than they need ;-) The dicey thing in all of this is that the generic kernels will be used for the certification of applications. If the cpu limit is too low then applications will simply not be certified for these high processor counts. One may encounter problems if the app is then run with a higher processor count. > > There are quite a few >16 socket boxes out there, which will give > > you >64 cpus with Montecito ... so I don't think that the >64 cpu > > system is going to remain in the noise for long. > > I bet the number of 32-way+ boxes is lost in the noise compared to the > number of 1-, 2- and 4-way boxes sold. Not that HP trust me with that > kind of sales data ;-) I think it is wrong to count the number of boxes. It is the number of users that is of major interest. These small boxes typically have one user. The larger boxes have whole communities working with them. The number of users of very large systems can potentially go into the 10 thousands of users per installation. Moreover these larger boxes significantly impact the technological progress of humanity on the planet. These large boxes run hurricane simulations, do complex visualization necessary for bio technological advances, cosmological simulations, other physic stuff etc etc. - 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/