Author: Robert Hyatt
Date: 14:12:17 01/06/00
Go up one level in this thread
On January 06, 2000 at 10:33:41, Graham Laight wrote: >On January 06, 2000 at 10:18:24, Robert Hyatt wrote: > >>>Probably even more important in Bob's case is that he has always wanted to >>>strongly promote the idea that multi-processor computers are massively stronger >>>than (single processor) PCs. It would be a blow to his esteem if it was proven >>>beyond all possible doubt that PCs had attained a GM level of play! >> >>pure nonsense... I believe multiprocessor systems are _better_. And that is >>provable by simple measurements. But it says nothing about how strong the >>things are on one cpu.. only that they are better on 2. > >Not quite pure nonsense. A multi-processor system is better than a single >processor system ONLY if they're both running the same software. > >What I was saying was that, because I believe that Bob has a strong emotional >preference for multi-processor chess systems over single (or small number) >processor chess systems (after all, he was the author of Cray Blitz, and he has >been known to strongly support DB in CCC threads), I believe that he would be >happier if a multi-processor machine became the first undisputed GM. This might >have an impact on his impartiality when judging whether PCs have reached that >standard of play yet. > >-g That is incorrect. I left supercomputers for a well-known reason: they are nearly impossible to get time on. I chose to go to a readily-available platform, the IBM PeeCee-compatibles, in order to have ready access to a machine whenever I wanted, without having to resort to dozens of phone calls and such. Yes, I would prefer that _my_ program became the first offical electronic GM. I doubt you would find _any_ programmer that doesn't feel the same way. But I have no interest in whether it will be a uniprocessor program or a multiprocessor program. I strongly suspect it will be a multiprocessor, simply because of horsepower. IE if I go to a drag race, I'm going with the biggest V8 I can find, not a 4-cylinder. Doesn't mean I don't pull for the 4 when he races other cars, of course. Remember also that my 'multiprocessor' machine is not horribly expensive. I put it together for about $12,000 total. Not cheap, but not outrageous. Bruce bought one himself and has it at his house. That is not exotic hardware by any measure today. A cray T932 for $60,000,000.00 is a different story.
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.