From owner-freebsd-stable Sat Jul 21 0: 8: 9 2001 Delivered-To: freebsd-stable@freebsd.org Received: from mail2.uniserve.com (mail2.uniserve.com [204.244.156.10]) by hub.freebsd.org (Postfix) with ESMTP id 4B9AE37B401 for ; Sat, 21 Jul 2001 00:08:07 -0700 (PDT) (envelope-from tom@uniserve.com) Received: from mail2.uniserve.com ([204.244.156.10]) by mail2.uniserve.com with esmtp (Exim 3.13 #1) id 15Nqrn-000IUq-00; Sat, 21 Jul 2001 00:07:59 -0700 Date: Sat, 21 Jul 2001 00:07:59 -0700 (PDT) From: Tom X-Sender: tom@athena.uniserve.ca To: Juha Saarinen Cc: 'Chan Tur Wei' , freebsd-stable@freebsd.org Subject: RE: FreeBSD 4.3 and 6G RAM In-Reply-To: <000901c111a2$f581e770$0a01a8c0@den2> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Sat, 21 Jul 2001, Juha Saarinen wrote: > :: Heh heh. Just a wild idea: How about a multi-processor > :: system FreeBSD where > :: each CPU gets mapped to different 4G segments? Perhaps some of the > :: address space can overlap to offer real SMP? > > Talking out of my hat here, but I don't think that's possible unless you > run a separate kernel on each CPU...? > > -- Juha Not really. The non-uniform-memory-access (NUMA) model used by SGI and IBM (Sequent) systems is pretty scalable. Each processor has dedicated fast memory, but can access other processor's memory space. SNP systems have a memory access bottleneck, that NUMA systems do not have, assuming your load can be balanced. Tom To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message