Author: Heiner Marxen
Date: 13:55:53 06/17/01
Go up one level in this thread
On June 17, 2001 at 15:38:50, Angrim wrote: >On June 16, 2001 at 08:30:19, leonid wrote: > >>Hi! >> >>I am not sure how deep this position is but it look like that only strong >>program will find it easy. >> >>[D]3K4/1Q1Q1Q2/Q1R1R1Q1/2N1N3/Q1qBNqQ1/q1qBRq1q/brnq1nrb/3kq3 w - - >> >>Please indicate your result. >> >>Thanks, >>Leonid. > >This sure has a lot of transpositions possible! >I just added support for transpositions to my pn^2 code, and with the >new code my chess position solver handled this nice and fast, but the >old version took longer than I have patience for. > >new version: >proved that move e4xc3 wins, 13 turns >PN2:12688384 evals, 286977 expands, 79.30 seconds > >I halted the old version after 190m evals, 1363 seconds. Wow! How interesting! From the thesis of Dennis Breuker I expect that adding transposition tables to PN^2 is not exactly trivial. Reading his thesis I have had several difficulties (some things look like plain wrong to me, but that can't be, so I just do not understand), and I would like to look at some working code, or see some pseudocode from someone with practical experience. So let me ask: what is the status of your code? May we look at it? Anyhow, thanks for sharing your results! Heiner
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.