Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: IBM Chess GM Chips

Author: Dann Corbit

Date: 15:09:28 03/30/99

Go up one level in this thread


On March 30, 1999 at 17:59:35, Peter McKenzie wrote:
[snip]
>I don't see why you are so dismissive of this.
>I'll just wait and see myself.
>
>It seems reasonable to believe that Hsu can make significant improvements to his
>chess chip and associated software.  For a start, chip fabrication techniques
>are constantly improving so there will of course be an 'automatic' speedup.
>
>I seem to remember that they hurried out a redesigned chip between the first and
>second Kasparov matches, so it is likely that a number of shortcuts were taken
>there to get it done in time.  Fixing these things, plus applying any leasons
>learnt when testing and tuning that chip, could conceivably result in a
>significant improvement.
>
>Don't forget that designing hardware is different to programming software.  In
>software, if you find a problem you can usually hack a fix pretty fast.  In
>hardware, you pretty much have to live with it or work around it until you can
>build your next prototype/version which may be some time off.  Hsu may have a
>good number of fixes stored up waiting to be implemented.
>
>I wouldn't write him off, he's done it once already :-)

I can tell you this, if it is under $1000 for the chess circuit board, I will be
in line to buy one [ I would probably have to talk to my wife about a $1000 toy,
of course ;-) ].

Every year, the trace size reduces for IC's.  Even without any improvement in
the design, the chips would be much faster just from a size reduction.  If they
can put multiple chips on a single small IC, they might produce some incredible
numbers.  The controlling software for these chips will be a new adventure,
perhaps.



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.