Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Has anyone compare Deeper Blue moves with their PC, 1 GHZ for 24 hours?

Author: stuart taylor

Date: 21:00:58 09/05/00

Go up one level in this thread


On September 05, 2000 at 21:47:03, pavel wrote:

>On September 05, 2000 at 20:59:04, Jonathan Lee wrote:
>
>>By using the latest Personal Computer at 1 GHZ, you could compare moves of
>>Deeper Blue (versus Kasparov) for "many hours".
>>What you do is put the same middle game position, and let your computer run for
>>hours to match the same move as IBM.
>>If there is a match in 24 hours, that would be 480 GHZ.
>>24 hours divided by 3 minutes per move equals 480.
>>Fritz 6 and the other contenders are recommended.
>>Jonathan (67th message)
>
>whose gonna waste so much time to play such a long game?
>atleast i wont.
>IMO its pointless and there is nothing to prove to anyone.
>besides you dont cmpare with DB with fritz or any other commercial chess
>softwares. ( no!! I am not saying DB is way too strong for all the others).
>
>this is because DB was a hardware solution (mainly). and most other commercial
>chess softwares are software solution.
>
>as i said there is nothing to prove ;)
>
>pavel


Why is there nothing to prove?  Hardware solution costs a good bit more than the
software solution, a few thousandfold, so if the software solution is looking
anything near to a similar level, then DB has to prove its money's worth. And as
long as that hardware (DB) is still thought to be that little bit stronger, it
just MAY be worth the extra few million (or many million).
 But even that looks very questionable, so we might prove that we don't need
the hardware solution.
  If DB is not "way too strong for all the others", then it might just be that
with the equivalent of 480ghz, the software might play greater chess, due to the
wisdom of the software programmers.
  Better still might be if they could make software to be maximized for 480ghz
use. At the momment, things are maximised for + - tournament timings on less
than a GHZ, therefore, much longer thinking time may not add very much
relatively. But who knows? it might add enough!
   I don't know if it is true, but CSTAL was claiming that with much more
powerful hardware, their program would get much better. If that is really true,
then perhaps 480GHZ with CSTAL might prove way above all other software on
anything, and also above Deeper Blue.
  Deeper Blue is/was maximised for all the power and speed given to it? Big
deal! I don't know if it was worthy of all that power. Maybe something maximised
for much less (even 480ghz) could do even better.
S.Taylor



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.