Author: blass uri
Date: 08:29:53 07/01/00
Go up one level in this thread
On July 01, 2000 at 11:03:18, Timothy J. Frohlick wrote: >On July 01, 2000 at 03:57:00, Jerry Adams wrote: > >> >> The Best Software on the best machines are not playing grandmanster level >>chess. Or should I say that Rebel century is an exception? Generally Software >>Programs in my opinion are not demonstrationing grandmaster results? I mean at >>40\2hrs or am I mistaken? > >Mr. Adams, > >I think that a Dr. Hyatt is pretty competent to make that assertion. He has >been doing this computer chess programming for over a quarter of a century. He >also knows about grandmaster level play. The machines just don't have a clue in >complex positions. They will get better. In five to seven years we will have 12 >teraflop home workstations and 8 man tablebase chess programs I guess that we are not going to have in 5-7 years 8 pieces tablebases. I guess that we will even not have the complete 7 pieces tablebases in 5-7 years. calculating the 5 piece tablebases can take at least some weeks and maybe more than a month in the hardare of today(calculating only one 5 piece simple tablebase with no pawns like KQR vs KR takes few hours on my pentiumIII450 and there are many tablebases to calculate when most of them takes more time). calculating all the 6 piece tablesbases will take more time in the hardware of 2007 without parallel search(I assume that computers are twice faster every 18 monthes so computers will be less than 30 times faster). parallel search can help but I do not believe it will be enough to generate all the 7 piece tablebases in 2007. There is also a problem of memory and I do not believe that there will be enough memory for the 8 piece tablebases in 2007 Uri
This page took 0.01 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.