Author: frank phillips
Date: 05:53:21 07/14/98
I have managed to write my first chess program with help from articles found on the Web and individuals in CCC (how to do the pv). Its a bit clunky (a lot clunky) but works, which was original aim. However it is disappointingly slow, managing only 4 to 5 ply (plus up to a 12 ply quiescent extension) in about 10-30 seconds or so in the opening on my K6-233, compared to Crafty, Exachess and GNU that manage 8-10ply or more easily in the same position with much more evaluation of leaf nodes. Any hints on how to make it go deeper, faster (on the same machine) would be most welcome ie the big areas to tackle. Reference to appropriate articles would also be welcome since I am new to programming, not just chess programming. The disparity in speeds is so great that it makes me think I could implement a technique or two to speed things up greatly (I hope) or have completely messed-up, which I doubt becuase it plays legal chess games (weakly) and solves a few WAC positions. The program does Negamax with capture extensions in the quiescent search, generates moves using the mailbox[] approach in Tom Kerrigan’s simple chess program, is based on integer arrays (not bitboard), stores moves for each node in a local array (of structures) in each incarnation of Negamax/Quiescent(), searches the pv move and captures first, does not have hash tables, does not do null-move, dose not do killer moves nor all the other things I have heard about but have no idea what they are. Frank BTW My thanks to Tom Kerrigan for his ‘simple chess program’ without which I would not have even got this far. Bob, if you ever write that chess computer programming book...............
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.