From owner-freebsd-hackers  Sat Oct 25 14:42:30 1997
Return-Path: <owner-freebsd-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id OAA25189
          for hackers-outgoing; Sat, 25 Oct 1997 14:42:30 -0700 (PDT)
          (envelope-from owner-freebsd-hackers)
Received: from isgate.is (isgate.is [193.4.58.51])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id OAA25180
          for <freebsd-hackers@FreeBSD.ORG>; Sat, 25 Oct 1997 14:42:27 -0700 (PDT)
          (envelope-from totii@est.is)
Received: from eh.est.is (eh.est.is [194.144.208.34]) by isgate.is (8.7.5-M/) with ESMTP id VAA01372; Sat, 25 Oct 1997 21:42:21 GMT
Received: from didda.est.is (totii@ppp-22.est.is [194.144.208.122])
	by eh.est.is (8.8.5/8.8.5) with SMTP id VAA24878;
	Sat, 25 Oct 1997 21:40:51 GMT
Message-ID: <345267BF.167EB0E7@est.is>
Date: Sat, 25 Oct 1997 21:42:23 +0000
From: =?iso-8859-1?Q?=DEor=F0ur?= Ivarsson <totii@est.is>
X-Mailer: Mozilla 3.01Gold (X11; I; FreeBSD 2.2.2-RELEASE i386)
MIME-Version: 1.0
To: wghhicks@ix.netcom.com
CC: freebsd-hackers@FreeBSD.ORG
Subject: Re: Parity Ram
References: <Pine.BSF.3.96.971025115335.173A-100000@trojanhorse.ml.org> <34524948.41C67EA6@est.is> <34525F3B.1137B612@ix.netcom.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by hub.freebsd.org id OAA25185
Sender: owner-freebsd-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

Jerry Hicks wrote:
> 
> Žoršur Ivarsson wrote:
> >
> > This has helped me several times when I was suspecting broken memory in
> > the old days (90-93) :-)
> >
> > Thordur Ivarsson
> 
> ECC Memory was marginally useful for this years ago when were using NMOS
> RAM. Lately, most memory failures I've seen are catastrophic, taking out
> a whole device or better.
> 
> I'm not a hardware specialist; Does 'Parity RAM' employ a conventional
> parity scheme, a la asynch serial communications?
> 
> Didn't Richard Hamming show these to -cause- more problems than they
> solve? It seems I recall a number like 256K (bits/bytes/words?) as being
> the threshold in a proof he presented.
> 
> Jerry Hicks
> jerry_hicks@bigfoot.com

In the old days 8088, 8086, 80186, 80286 and 80386 it was just plain odd
or even
parity like in serial communications but later on came better algorithm
that could
really fix wrong bits

Thordur Ivarsson