Author: Uri Blass
Date: 09:20:00 04/21/01
Suppose the processors are limited to analyze chess position in the same way that the main program is doing it . It means that you can calculate only score,move and all the main lines when you get information about the board position and the time to search. What is the maximal speed improvement that you can get by using x processors. It is clear that you can get some speed improvement even from 2 processors because at the same time that one of the processors is playing another processor may ponder and if it is lucky to ponder on the right position(it can use the main line to increase the chance to be lucky) it can use the information later on the game. Another point is that 2 processors may ponder on different moves and again the changes in the main line may help(if the main line was changed from 1.e4 e5 to 1.e4 c5 in the last ply then it may be logical to ponder on both 1.e4 e5 and 1.e4 c5). Uri
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.