Author: Dann Corbit
Date: 11:11:15 05/13/99
Go up one level in this thread
On May 13, 1999 at 11:34:23, Chris Carson wrote:
>Potential Advosary: DB the chip
> 30M NPS single chip, up to 4 supported
> estimated price $200 single chip
> strength: strong GM at tournament time controls
>
>How do the win/tel programs strike back?
>
>Let's discuss some proposals. Here are couple of my
>suggestions:
>
>1. Create xxxx the chip. where xxxx is Fritz, Hiarcs, Rebel, Crafty, ...
>2. Multi-processor versions of current programs.
>3. Hybrid solutions - chip accelerators, multiprocessor, ...
>4. Play on weakness (if any), ie- DB the chip may be more expensive
> than $200.00, or DB the chip may not be as fast/strong as
> expected (DB beat commercials 38:2, but commercials were running
> on pentium 90 machines, both s/w and h/w will be much faster/
> stronger when DB the chip is released than the P90.) also, once
> DB the chip is on the market, everyone will get a chance to play/
> analyze and exploit it's weakness, every program/machine has
> some. :)
>
>I am sure we can come up with a few more suggestions. Ok, take pot
>shots at my suggestions, but I hope the commercial programmers are
>reading this. HSU is a great scientist, but he is not the only person
>that can create a hardware solution. :)
How many dollars do you imagine went into the creation of deep blue from the
time it was a university project up till the time it beat Kasparov? I imagine
that quite a bit has been done since then.
I'd be shocked if anyone could beat them for less than 50 million dollars [just
a guess, but years {at least a decade -- anybody know exactly how long?} of
university research and corporate sponsorship can amount to quite a pile of
dough]. That's a pretty sizeable head start. And they have already worked out
the glitches. So your tools had better be pretty well done if they are to have
any chance of competing. If it does come to pass, it will be the only game in
town. Unless you are not willing to pay what it costs, you will simply have an
inferior product if you use anything else.
On the other hand, the chips were controlled with C language programs -- it was
not a 100% hardware solution. I suspect that there are programming challenges
still awaiting. But the current metaphors would have to be totally rethought.
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.