Author: James T. Walker
Date: 10:40:21 04/01/99
Go up one level in this thread
On March 31, 1999 at 08:27:52, Harald Faber wrote: >On March 31, 1999 at 07:28:47, Dave Gomboc wrote: > >>>>>> The last paragraph states that he is forming an independent start-up >>>>>>company to create a new chess chip for consumers , that would make it possible >>>>>>for a desktop machine to defeat the world champ in a formal match by 2000 .... >>>>> >>>>>That confirms my assumption that Hsu has gone ***... >>>>>A chess chip becoming WC in 2000....nonsense. >>>> >>>>The quote says "a desktop machine". It does not say "one chess chip". I'd >>>>expect there to be many chess chips in one machine. Total hand-wave here: if >>>>there were 16 chips per PCI card, and 4 PCI slots on a mainboard, that'd be 64 >>>>chess processors in all. Now if a (Dual? Quad? More? :-) Pentium III can keep >>>>them busy, I can readily imagine such a machine opening up a can of serious >>>>whoop-ass. :) >>>>Dave Gomboc >>> >>>Maybe I am wrong. >>>But I don't believe Hsu nor anyone else being able to CREATE a chess chip (read >>>above), build several of them on a special board maybe and write a software that >>>a) handles these multiprocessors with appropriate tasks and b) becomes WC, all >>>within one year. >>>Just my opinion. >> >>Hsu has already designed two chips, maybe he's got incremental adjustments for >>the third in mind, or maybe he's just planning on using the second design. >> >>IBM has fabricated the chips before. Seeing as Hsu worked for IBM for many >>years, it seems reasonable to me that they could work out some kind of >>arrangement. >> >>As for "special boards", well, they've done this already too. I think they are >>VME-bus, but it can't be that hard to use PCI instead. >> >>What's left? Parallel search? Not a problem. They already do that. > >I didn't mean that they are not able to do that. > >>As for becoming the world champion on a desktop machine, I'll agree with you >>there, that sounds like hyperbole to me. And you're right, these things will >>all take time. Maybe 2000 is doable, but probably not January 2000. :-) >>Dave Gomboc > >That is the point. They may be able to develop such things but this won't be >enough to become WCH next year. I believe this is all possible. They probably have a very good start on the chip (Deeper Blue chips). They need to build the machines(Not that difficult). They need to get the World Champion to agree to a match. He is chomping at the bit to get revenge on anything associated with the Deeper Blue program/computer/team. He just might take a late Y2K match. Jim Walker
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.