Author: Dave Gomboc
Date: 09:06:13 03/23/00
Go up one level in this thread
On March 22, 2000 at 03:54:22, Thorsten Greiner wrote: >On March 21, 2000 at 12:54:16, Pete Galati wrote: > >>As it is now, and without adding Dann's pgn files yet, a 13.2mb pgn compiles a >>20mb book with Amy, so I expect it to end up being on the large size before I'm >>done, I'm hoping Amy doesn't have a limit. Amy's a cool little program. > >Pete, > >Amy internally uses an open source but commercial grade database engine called >Berkeley DB (see http://www.sleepycat.com). I am not aware of any size >restrictions Berkeley DB has, other than those of the underlying operation >system (many OSes limit file sizes to 2 GB). > >To index the database Amy uses a 64 bit integer which is computed using the >usual Zobrist hashing function. There is a very tiny chance that to >non-identical positions hash to the same 64 bit integer, so that a bad opening >move might be played. But I do not consider this to be relevant in real life. > >In the 0.7 version of Amy which I will release shortly the book size is >decreased a little bit. Also you will have the option to 'flatten' (is this a >good english word for what I mean) "prune" is the word you want here > the opening book by removing positions which >occur less than 'n' times. Even for n=1 I found that opening books reduce to >1/10th of the original size. Why are there positions in your book that occur zero times? (Or did you mean "removing positions that occur less than or equal to 'n' times"?) >Hope you enjoy Amy > > -Thorsten Dave
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.