Return-Path: <John.Harrison-request@cl.cam.ac.uk>
Delivery-Date: 
Received: from leopard.cs.byu.edu (no rfc931) by swan.cl.cam.ac.uk 
          with SMTP (PP-6.5) outside ac.uk; Tue, 29 Nov 1994 17:46:30 +0000
Received: by leopard.cs.byu.edu (1.38.193.4/16.2) id AA16233;
          Tue, 29 Nov 1994 10:38:39 -0700
Sender: info-hol-request@lal.cs.byu.edu
Errors-To: info-hol-request@lal.cs.byu.edu
Precedence: bulk
Received: from bilbo.radonc.washington.edu by leopard.cs.byu.edu 
          with SMTP (1.38.193.4/16.2) id AA16229;
          Tue, 29 Nov 1994 10:38:38 -0700
Received: by violin1.radonc.washington.edu (1.38.193.4/16.2) id AA14454;
          Tue, 29 Nov 1994 09:31:19 -0800
Date: Tue, 29 Nov 1994 09:31:19 -0800
From: Jon Jacky <jon@violin1.radonc.washington.edu>
To: info-hol@leopard.cs.byu.edu
Subject: Re: Pentium bug --- algorithm or hardware?
Message-ID: <"swan.cl.cam.:056310:941129174716"@cl.cam.ac.uk>


John Harrison asks:

> I'm curious whether the problem was in the underlying algorithm used
> or its hardware implementation ... it may be that formal
> verification wouldn't have helped.

I'm curious about this too.  It really isn't possible to tell from the
news stories we've seen so far.  I recall hearing that similar
difficulties with other chips have been due to very low-level
electric/materials problems, not logical problems: charge was slopping
around on the chip when particular bit patterns were present, and they
fixed it by changing the chip geometry to make some traces and pads
larger.  (I can't recall any reference for this ... it may be just a
rumor ...  I seem to recall hearing that some of those low-level
problems were even sensitive to the operating temperature).

- Jon