Author: Robert Hyatt
Date: 11:17:24 01/26/98
Go up one level in this thread
On January 26, 1998 at 13:13:38, Johanes Suhardjo wrote: >First of all, sorry for the slow response. I don't have a real internet >access in the week end. > >Bob is right, I got the wrong position. However, I should mention that >both crafty 's "setboard" and my "fen" commands give White King at a2 >when >parsing 8/k/3p/p2P1p/P2P1P/16/K w - - 0 1. > ^^ I didn't notice that. 16 is illegal in fen. There should be no more than 8 values between each pair of "/" characters. So this should be interpreted as 1 empty square followed by 6 empty squares, which is wrong. >Both commands get the right board after 8/k/3p/p2P1p/P2P1P/8/8/K w - - 0 >1. > >With the correct board, crafty-11.13 solved it in 19 plies in 9 seconds, >crafty-14.4 in 19 plies in .37 seconds. Both gnuchess-4.75 and my >program >got stuck at ply 19. Gnuchess got there after 1136132 nodes in 62 >seconds >and my program got there after 24920 nodes in 2 seconds. Not >surprising, >my program is based on an older version of gnuchess, so both are likely >to >have the same problems. > >Since Bruce said it is likely a "horrible bug", I think I will spend >some >time fixing it before playing with other ideas. With this wonderful >Computer-Chess Club, I'll appreciate any help from you gurus. > > >On January 23, 1998 at 22:31:22, Stuart Cracraft wrote: >>I think you're having a problem other than the number of slots. >>What is your replacement strategy for how table entries get >>overwritten? > >It rehashes linearly up to the next 6 slots. If there's a conflict, the >entry with deeper search is kept, if the depth is the same, the newer >entry >is kept where the age of an entry is determined by the node count. > > > Johanes Suhardjo (johanes@farida.cc.nd.edu) >-- >The older a man gets, the farther he had to walk to school as a boy.
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.