Author: pavel
Date: 08:48:24 09/20/00
Go up one level in this thread
On September 20, 2000 at 05:48:47, Keith Kitson wrote: >I would expect a large position learning database, which continues to grow will >use more of the processor time to search this growing database to a point where >precious little time remains for the processor the execute much of the logic in >the program for any given move. Plus as this learning db will be so large it >will have to reside on disc, and so access times will be slower than hash >memory. > >It doesn't seem to be the right way to go in the short term. if chess programs can use TB, which may contain 20gb in size, then why not chess database? the main problem is to find a way to port the program in such a way that it will search the database as it searches TB. some are saying (which is true) that hash file plays an important role and thus a perticular position with PV wont do any good without hash. now the question is which part of the hash to be stored? the hash of the whole game?.......it might turn out to be reaalllly big. any other options? Pavel
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.