Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hardware and WCCC limits?

Author: Robert Hyatt

Date: 11:25:12 05/19/04

Go up one level in this thread


On May 19, 2004 at 11:59:19, Uri Blass wrote:

>On May 19, 2004 at 11:24:41, Vincent Diepeveen wrote:
>
>>On May 19, 2004 at 10:38:08, Matthew Hull wrote:
>>
>>>On May 19, 2004 at 10:30:22, Vincent Diepeveen wrote:
>>>
>>>>On May 19, 2004 at 10:25:05, Matthew Hull wrote:
>>>>
>>>>>On May 19, 2004 at 10:09:29, Vincent Diepeveen wrote:
>>>>>
>>>>>>On May 19, 2004 at 09:57:04, Daniel Clausen wrote:
>>>>>>
>>>>>>>On May 19, 2004 at 07:27:00, Vincent Diepeveen wrote:
>>>>>>>
>>>>>>>>On May 19, 2004 at 03:35:39, Daniel Clausen wrote:
>>>>>>>>
>>>>>>>>>On May 19, 2004 at 03:02:05, Russell Reagan wrote:
>>>>>>>>>
>>>>>>>>>[snip]
>>>>>>>>>
>>>>>>>>>>That is the goal of the WCCC, to have a competition between the best
>>>>>>>>>>computerized chess playing entities, not to test which software is the
>>>>>best. >>>>
>>>>>>>>>Yes, but strictly speaking this would mean that a company selling program
>>>>>XYZ >>>>wouldn't be allowed to write WCCC2004 winner on their software
>>>package, >>as it >>>>was not the software but software+hardware which actually
>>>won the >>tournament. >>>>
>>>>>>>>>Sargon
>>>>>>>>
>>>>>>>>This is a nonsense of course.
>>>>>>>
>>>>>>>Let me rephrase it: when you play a tournament like WCCC, where the pair
>>>>>HW/SW >>is tested and your particular HW/SW combination wins, then it's not
>>>>>"correct" to >>silently skip the HW part and just claim your SW part won.
>>>>>>>
>>>>>>>I guess you still think it's nonsense. If so, you're welcome to be a bit
>>>more >>>>specific and say _what exactly_ is nonsense about it.
>>>>>>>
>>>>>>>Sargon
>>>>>>
>>>>>>You really are overrating hardware.
>>>>>
>>>>>
>>>>>This is uniquely true of DIEP.  No matter if it runs on a 90 mhz machine or a
>>>>>400cpu x 500mhz supercomputer, it plays consistently at the FM level -- a
>>>feat >>of programming mastery unequaled in the annals of computer chess.
>>>>>
>>>>>However, it would be a mistake to assume other projects are as well
>>>engineered. >
>>>>Why not show up at world champs operating crafty?
>>>
>>>
>>>Crafty does not need an operator.  Crafty incorporates features of automation
>>>which permit it to play entire tournaments without human intervention.  Since
>>>Omid says that internet access is available at the event, all you have to do is
>>>connect to ICC with free software.  Then you and the other participants will
>>>have the chance play the winner of the latest Internet World Computer Chess
>>>Championship.
>>>
>>>BTW, where was DIEP at the last IWCCC?  I guess the programmer didn't think he
>>>could win.
>>
>>At saturday i played a game for my team (www.schaakclubutrecht.nl). A fide rated
>>game 40 in 2 + 20 in 1 + 30, but i'm sure you have not a f'ing idea what i talk
>>about here, and at sunday i played for my belgium club Hoboken at 40 in 2 + 1
>>(also FIDE rated of course, all important leagues are FIDE rated).
>>
>>So i had no time for a small internet tournament where the level is not even
>>serious. 45 10 at my dual k7 means after a few moves like 30 seconds a move.
>>
>>This at 2 x 2.127Ghz. Let's say that's 0.5 * 2 * 2127 Mhz = 2127 Mhz minute.
>>
>>At world champs 1999 the chessbase top played at 4x500Mhz at 3 minutes a move.
>>So that's 6000Mhz minute.
>>
>>How can one take an internet tournament serious where the level+hardware is for
>>sure slower than what was the standard in 1999?
>>
>>Also if there is commercial interests in a tournament at the internet, then
>>curious things will happen. Additionally there is hardly programmers interesting
>>enough to chat to at such CCT tournament.
>>
>>So it is good that i play real tournaments, and yes it doesn't matter whether
>>world champs plays in USA or in Europe or in Asia or in middle east. In
>>principle i will come and join.
>>
>>And no i would not have a problem with a world champs in UAE either nor USA. I
>>would welcome it. I wonder though whether it political would be correct to
>>organize something in USA nowadays.
>>
>>People like you should stop having a big mouth about ICGA organizers.
>>
>>It is true they ask money to organize an event and it is true that i am not a
>>big fan of the amount of money they need to organize an event.
>>
>>However US organisations in the past needed a 10 fold budget to organize events
>>there, so just do the math...
>>
>>In fact you can play a lot of FIDE rated events here with entry fees ranging
>>from 40 euro to 125 euro. That's 9+ rounds or so.
>>
>>In USA a default participation in a tournament is like 250 dollars a person.
>>
>>Incredible prices.
>>
>>And you wonder why there is no computerchess tournaments organized in USA?
>>
>>That's because your big pal Hyatt tells at radio interviews that nothing ever
>>will be able to equal deep blue. So no company of course is interested in USA to
>>invest in computer chess world champs nowadays.
>
>Hyatt never said that nothing ever will be able to equal deep blue.
>When do you plan to stop posting lies about him.
>
>Uri


Never.  It is all he actually knows how to do...

Just like I supposedly said that my generalized speedup approximation formula
works for _any_ number of processors when I have _always_ limited that formula
to hardware I have actually run on...

Just like I supposedly wrote a JICCA paper about the Crafty parallel search.  I
plan to do so, but it hasn't been written or published yet.  Yet he claimed it
had, and then accused me of denying something I had written.  :)

We should have a CCC poll question "Does _anybody_ believe that Vincent is a
credible person regarding claims he makes here?"

I suspect it would get 1-2-3 "yes" votes and that would be all...

For obvious reasons...



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.