Author: Adrien Regimbald
Date: 19:40:42 05/12/00
Hello, I am working on implementing a binary book into Faile. I am trying to get some rough figures on certain things to have an idea of how much memory I will need for book creation, as well as how big the output binary file would be. Does anybody have any idea about figures for the following: - average number of new positions found in a new game at various book ply - 20, 40, 60 ply, for example - needed hash entries per unique opening position (ex - say I compute I will get about 50,000 unique positions from my input PGN file, and want to store them in a hash table as I generate the book, how many hash entries would I need to avoid collision problems?) - how much a binary book file's size could be reduced by using an array of found positions rather than a raw dump of the hash table you used to generate it - gain in unique position count vs. game count (ex - if I have a 10,000 game book, and it gives say .. 20,000 (guessing) unique positions, how many should I expect from a 1,000,000 game book) Any guidelines anybody can give me concerning this subject would be greatly appreciated! Regards, Adrien.
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.