Author: Robert Hyatt
Date: 20:06:08 08/16/03
Go up one level in this thread
On August 16, 2003 at 21:04:45, Federico Corigliano wrote: >Some chess engines (mainly bitboard engines as Crafty and Pepito), in the moment >of the recursive search, don't generate all the moves in one step, first try the >hash move, then generate the captures, and if there aren't FH, follow with the >killers, and so on. This method seems that is good for some speed increment. >In my engine I have hash table move, captures ordered by SEE, three killers, and >history heuristic, and I generate all the moves at the same time. How much can >be speed increment if I split the move generation? > >Greetings and sorry for my cloudy english :-) >Federico It turns into a significant improvement, because of the q-search. IE if you can generate _only_ captures/promotion moves, then you won't have to pass over the non-captures while trying to select good captures...
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.