Author: Dann Corbit
Date: 21:26:25 02/02/99
Go up one level in this thread
On February 02, 1999 at 23:50:19, Paul Richards wrote: >On February 02, 1999 at 15:29:48, Terry Presgrove wrote: > >>So they have the technology in place to produce the 2 billion nps >>but not the actual hardware manifested in a resurrected Deep Blue. >>Reading between the lines that was my (limited) understanding also. > >Yes it's marketing, so was the original match. However I still >think it is an interesting question, how strong a Deep Blue could >be built today? Is 2 billion nps correct assuming same number of >processor nodes? What is that then, 20 times faster? While I >think Kasparov became fatigued, and also was somewhat handicapped >by not being able to study prior games of the machine, I think >even he would hesitate to accept a match with a machine 20X more >powerful than the one he faced last time. If such a machine were >to enter regular competition, I think that contrary to Kasparov's >claim that he would crush it, it would crush him and all others. >It might lose here and there until a few adjustments are made, >but if it were an all out contest where the team refines the machine >after each tournament, it would easily be the World Champion in >short order, if not right out of the box. I also think that the >reason that IBM might not want to do such a thing is not because >it fears a loss to Kasparov and the marketing result, but rather >if the machine became a regular competitor it would have to >accept challenges from other supercomputers, and THAT is what would >be frightening from a marketing standpoint. 20x faster gets you about one more ply, on average (1/2 fullmove) in the same time frame. I don't think that aspect of the machine would be particularly that much more frightening because of the exponential nature of chess. However, some of the new technology like Eugene's tablebase generator could be used to create 6 piece tablebase files. So you had better wrap it up early, because in the late endgame there will not be flaws in play. Every new invention and technical paper on chess algorithms is also available to the IBM team to innovate from. While one ply deeper does make a truly fearsome opponent somewhat more fearsome, it is the combination of all the advancements available to them that would make the next generation even more formidable.
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.