Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Off Topic: IBM claims breakthrough in processor technology

Author: leonid

Date: 16:38:33 03/03/00

Go up one level in this thread


On March 03, 2000 at 14:43:51, Tom Kerrigan wrote:

>On March 03, 2000 at 06:46:35, leonid wrote:
>
>>It sound to me otherwise. If hardware will be from one day to the next one more
>>speeded by factor of 1000 or 10000 everything will be changed dramatically. For
>>now we must wait more that 1 year for some doubling of the computer speed.
>
>Yes, but if you read the article again carefully, you will see that they are not
>claiming an immediate speedup of 10000. Maybe a speedup of 3 in several years,
>which just continues Moore's law.


If this is expected speed it is just nothing. Nothing to dream about.

>
>>approach could replate many advanced pruning technincs. If, for instance, at the
>>end of the game, when we have already only few pieces on the board, computer in
>>natural way will be able to see around 18 plies deep, there is the possibility
>>that all the database for the end of the game could be forgotten. It is normal
>
>That _might_ be reasonable if endgame databases are only probed at the root of
>the tree. But most programs probe them during the search, so they should
>definitely not be taken out simply because computers are getting faster. If
>anything, they become more useful as computers get faster.


Only at certain moment all those "extras" will be nothing but "overkill".


>>to expect that speed of thinking and precision of chess game run on PC will
>>overwhelm for ever every human in this Planet. At that time the best chess game
>>will be not the game that play better game (all of them will be better that
>>human) but those that are only bug free.
>
>I disagree. You have to remember that processors today are thousands of times
>faster than processors 20 years ago, but that doesn't trivialize the importance
>of a good algorithm. TSCP is pretty bug-free, but that doesn't mean it can
>compete with, say, Junior.
>
>-Tom


20 years ago there was no program able to beat decently playing human adversary.
Now even starting program can humiliate 95% of normal players at high speed
game.

Leonid.



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.