Author: Anthony Bailey
Date: 12:21:21 09/19/99
Go up one level in this thread
Eugene Nalimov: >I am really using one byte per position. And yes, 4Gb is overkill - 2Gb will be >more than enough. It just happened that I could use 4Gb machine, and had no 2Gb >machines... Ah, how frustrating that must have been for you. (c: Thanks very much for indulging me with the basic tablebase generation algorithm; I appreciate it. >Please note that algorithm is not the best one - it is the simplest one to write >and debug, so I could concentrate on more interesting/important parts of the >program, and I had enough horsepower to use it. With regard to better algorithms; I'm interested to know whether ones that work by "unmoving" pieces back from evaluated positions rather than repeatedly moving them forwards from unevaluated positions have been given attention. (I would assume so; given that the average number of ply required to reduce a position is L it would seem to reduce the number of times you have to calculate the legal moves (or "unmoves") in any one position from L to 2. I suspect this is all old ground, but I'm uncertain as to where the best place is to go to catch up; if there are any references available on-line, I'd love to know about them.) - Anthony.
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.