Author: Uri Blass
Date: 12:04:21 06/20/01
Go up one level in this thread
On June 20, 2001 at 14:43:57, Scott Gasch wrote: >On June 20, 2001 at 10:34:01, Robert Hyatt wrote: >> >>I obviously didn't use it. :) I added that because Peter Mckenzie posted some >>screwey position a few years ago and asked for those numbers. His position had >>most everything possible very quickly, castling, enpassant, promotions, >>you-name-it. >> >>You are debugging this about the only way you can. However, you are making one >>mistake that might well cost you a lot of time. What happens if _I_ have a bug >>and produce the wrong number? :) > >If you post the position, Uri, the rest of us with plytesting commands can run >it also and give results... > >Scott The position that Bob meaqnt was posted in http://www.icdchess.com/forums/1/message.shtml?176034 I only tested the initial position for debugging and did not test the position in that mesage. I am going to try to see if there are bugs for the initial position at 6 plies and only if there are no bugs I am going to try the relevant position. My program is very slow and needs more than 4 minutes to calculate perft 5 and more than an hour to calculate perft 6 for the initial position. One of the reason is the fact that I did not use optimize for speed in my project(I work with visual C++6) but the main reason is probably different. I know that I can improve the speed of my program but I prefer first to make it work and only later try to improve the speed. It is clear that I can improve the speed by using the last move to update my squaresee,directionsee and pin array(I now use only the squares of the pieces in the board) Uri
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.