Author: Robert Hyatt
Date: 16:54:30 11/08/99
Go up one level in this thread
On November 08, 1999 at 18:45:57, Alexander Kure wrote: >On November 08, 1999 at 17:35:11, Robert Hyatt wrote: > >>On November 08, 1999 at 14:27:35, Alexander Kure wrote: >> >>>On November 07, 1999 at 21:09:04, Robert Hyatt wrote: >>> >>>[snipped] >>> >>>>Lang may have dominated the micro programs.. but he _never_ dominated computer >>>>chess. The 'program to beat' went like this: >>>> >>>>1960-1970 MacHack (Greenblatt) >>>>1970-1977 chess x.x (slate) >>>>1977-1979 chess x.x and belle (slate/thompson) >>>>1980-1982 Belle/Chess x.x/Cray Blitz (slate, thompson, hyatt) >>>>1983-1986 Cray Blitz >>>>1987-present deep thought/deep blue (Hsu) >>>> >>>>No other programs were close during those time periods, if you talk about >>>>'micro programs'. >>> >>>[snipped] >>> >>> >>>Hi Bob, >>> >>>I think that your last sentence neglects the fact that Fritz 3 running on a >>>Pentium 90 MhZ beat Deep Thought in Hongkong 1995. After this 'disgraceful' >>>event the micros took the lead over the mainframes. >>> >>>Greetings >>>Alex >> >> >>You _really_ believe that? They lost two whole games to other computer >>programs during a 12 year span of time, and they were 'taken over'??? >> >>I wish you had a chance to try on Cray Blitz at 7M nodes per second. You >>might discover that it is _not_ exactly a patzer. And it isn't close to >>deep blue either... based on games _actually_ played vs them. For every >>micro win over a 'mainframe' someone can dredge up 10 losses to mainframes. > > >I would be interested to see some of these losses of recent date ;-) > The last even they played in was Cape May in 1995. They blew everyone out. I know of only two losses they had from 1987 on... one in 1995 to Fritz (Hong Kong) and one other that slips my mind at the moment. That really is a remarkable performance, not matched by chess 4.x, belle, Cray Blitz or anybody else... really remarkable... > >>I don't think the gap has closed at all... it has spread further, because the >>micro computers of today are _nowhere_ near the supercomputers of 5 years ago. >>In raw computing speed or any other measure... >> >>And the micros aren't even in the same rating pool with deep blue. > > >Maybe the fact that Micros did not have a chance to win more often than 2 whole >games (why only 2? which one besides Fritz vs. Deep Tought?) during the last 12 >years was due to the lack of possibilities? After the Hong Kong fiasko Deep >Thought did not participate in any computer chess tournaments any longer (if I >am not mistaken, please correct me if I am wrong). They participated in every ACM event that I remember. I hardly missed a one from when they first showed up with a decent version in 1987 (1986 they were a 'bust' with many hardware problems). So they averaged 4-5 games every year at ACM events, for 12 years.. that is about 50 games, not counting the WCCC events every three years also, which included 1989 and 1995 (they did not go in 1992). >And this year in Paderborn I wondered where the powerful supercomputers were? >The ones who showed up did not really play an impressive tournament (maybe >Cilkchess just did well enough). >I am sure there must be some big shots (besides Deep Blue who no longer exists >and therefore is irrelevant to take into account any longer) out there but my - >maybe false - impression is that they are beyond recognition. >I think the gap between Micros and the so called supercomputers has not spread >any further since Honkong 95, on the contrary I would say the Micros are >beginning to close the gap. You simply don't understand the hardware then. IE on 1995 hardware CB was hitting 7M nodes per second. I haven't tried new hardware from Cray because I got tired of the scheduling problems. But crafty can exceed that on the right alpha. IE on current 21264's at 750mhz a single cpu version can hit over 1M nodes per second. on a 16+ cpu machine it would be a _very_ difficult opponent for anyone to beat. And that doesn't get close to deep blue at all. They are simply impossible to handle at present, and their current hardware won't be touchable for at least another 10 years, if not longer. > >Greetings >Alex
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.