Author: Christophe Theron
Date: 22:07:43 04/11/03
Go up one level in this thread
On April 11, 2003 at 13:37:37, Peter Berger wrote:
>On April 11, 2003 at 13:14:21, Uri Blass wrote:
>
>>On April 11, 2003 at 13:02:19, Peter Berger wrote:
>>
>>>On April 11, 2003 at 12:20:31, Christophe Theron wrote:
>>>
>>>>>That's a silly stance to take. If you look at the progress of computer chess
>>>>>over the last 10 years I would wager the *vast* majority of the progress can be
>>>>>attributed to speedups in hardware. Simple test - take a 10-year old chess
>>>>>program and run it on today's hardware and see what it's strength is. (OK so
>>>>>maybe I'm overstating things here - this would be an interesting test, has it
>>>>>been done?)
>>>>
>>>>
>>>>
>>>>On equal hardware the programs of 10 years ago get slaughtered.
>>>
>>>This is not true at the very top IMHO if you make the comparison on 10 year old
>>>hardware. (I hope Genius 3 is 10 years old already :) ).
>>>
>>>On a 486DX50 Genius still is a powerhouse.
>>>
>>>Peter
>>
>>Does it have chances against the comercial of today in 120/40
>>when you use 486?
>
>I can only guess based on games/testresults I observed at faster timecontrols
>(g/30, g/60) that gave me the impression it can. I haven't let them play games
>at 120/40. Also there were not enough games.
>My impression is based on testgames against Chess Tiger for the Palm.
>
>>
>>Do you believe that the main disadvantage of Genius is earning
>>less nodes from better hardware so 120/40 on 486 is not eqvivalent
>>to 5/40 on faster hardware.
>>
>
>I believe that 120/40 on 486 is equivalent to 5/40 on faster hardware but a 486
>is a very slow computer and I suspect the computer where I could let it play
>5/40 would still be quite slow - it's only factor 24 between the time controls.
>
>On a P90 Tiger 11.8 was stronger than Genius at 120/40 (SSDF) but I don't know
>much faster a P90 is. I calculated it once but forgot the result :).
>
>I have a P233 though - if someone can calculate how much faster it is for
>chessprograms I could play some matches.
>
>Peter
You are right on this point, Peter.
It is true that Genius on slow hardware is still today playing at a decent
level.
However you really have to use slow hardware and fast time controls.
I remember two years ago playing blitz games Tiger 14 against Genius 5 on a K6-2
450MHz, and Genius was clearly beaten (I think the score was 75% in favor of
Tiger).
When I did the same experiment on 386sx-20MHz, Genius was doing much better,
actually winning the match (but I do not remember the winning percentage).
Christophe
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.