From owner-freebsd-alpha Wed Apr 3 9:37:51 2002 Delivered-To: freebsd-alpha@freebsd.org Received: from freebie.xs4all.nl (freebie.xs4all.nl [213.84.32.253]) by hub.freebsd.org (Postfix) with ESMTP id 37F7837B41C for ; Wed, 3 Apr 2002 09:37:45 -0800 (PST) Received: (from wkb@localhost) by freebie.xs4all.nl (8.11.6/8.11.6) id g33HbYP48093; Wed, 3 Apr 2002 19:37:34 +0200 (CEST) (envelope-from wkb) Date: Wed, 3 Apr 2002 19:37:34 +0200 From: Wilko Bulte To: Bernd Walter Cc: Christian Weisgerber , freebsd-alpha@FreeBSD.ORG Subject: Re: Source of "processor correctable error"? Message-ID: <20020403193734.A48068@freebie.xs4all.nl> References: <20020401152244.GE41357@cicely8.cicely.de> <20020402163220.GO41357@cicely8.cicely.de> <20020402192425.B43540@freebie.xs4all.nl> <20020402232244.GP41357@cicely8.cicely.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20020402232244.GP41357@cicely8.cicely.de>; from ticso@cicely8.cicely.de on Wed, Apr 03, 2002 at 01:22:45AM +0200 X-OS: FreeBSD 4.5-STABLE X-PGP: finger wilko@freebsd.org Sender: owner-freebsd-alpha@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On Wed, Apr 03, 2002 at 01:22:45AM +0200, Bernd Walter wrote: > On Tue, Apr 02, 2002 at 07:24:25PM +0200, Wilko Bulte wrote: > > On Tue, Apr 02, 2002 at 06:32:20PM +0200, Bernd Walter wrote: > > > Generaly speaking they have to be the same because the are all > > > addressed in a single bank and so claims the documentation. > > > But practical speaking you could try a random vendor 24 chip simm > > > with a very good success chance. > > > I wouldn't try to mix with a 18 chip simm. > > > Theoreticaly even edo should work, but you don't get a speed gain > > > and they are rare. > > > > Are you absolutely sure? I don't think we ever used EDO *SIMMs* > > on Alpha machines. > > EDO and fastpage are nearly the same. Nearly.. One of my direct colleagues was one of the folks who was involved with qualifying RAM/simm vendors for Alphas. I've heared enough stories of junk RAM to be very weary of 'nearly the same' ;) > EDO only has a longer data output phase while traditional fast page > go into tri-state. > The speed gain is that you can already take the steps for the next > access, while still taking the current data. > If you have a board not doing bank interleaving they are often > interchangeable as the small difference doesn't matter. There is of course the matter that you need to find 36 bit EDO SIMMs to be able to test this in the first place. EDO PC stuff tends to be 32 bit only in my experience at least. > > There are some boxes using EDO *DIMMs* (like the AS500 > > sitting next to me) > > Nice shoot - it's the same chipset (21172) as the PC164... > But no sign of a possible speed gain in the chipset docs. > They should work with identic speed as fastpage. You can try of course. -- | / o / /_ _ wilko@FreeBSD.org |/|/ / / /( (_) Bulte Arnhem, the Netherlands To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message