Author: Uri Blass
Date: 00:36:35 12/09/01
Go up one level in this thread
On December 08, 2001 at 20:08:03, leonid wrote: >On December 08, 2001 at 17:46:49, Uri Blass wrote: > >>On December 08, 2001 at 17:23:06, leonid wrote: >> >>>On December 08, 2001 at 15:40:29, Heiner Marxen wrote: >>> >>>>On December 08, 2001 at 10:27:55, leonid wrote: >>>> >>>>>[D]2R2R2/Q6Q/B1nqqn1B/N1qkqp1N/Q1qqqp1Q/Q1rrbb1K/Q6Q/2N2N2 w - - >>>>> >>>>>Please indicate your result. >>>>> >>>>>Thanks, >>>>>Leonid. >>>> >>>>Chest says: No solution in 9 moves. K7/600, 350MB, 2.9 hours. >>>>Since the EBF is around 10, I stopped here. >>> >>> >>>Thanks, Heiner! >>> >>>Now I know that this is mate in 10. Could hardly expect to see 9 moves deep >>>since my 8 moves, by brute force, took already 1 hours and 53 sec. Mate was >>>found by selective in 10 moves. >>> >>>Cheers, >>>Leonid. >> >>Here are the results of my program on p800(perft 5 is the number of legal games >>of 5 plies) >>perft 4=15444557 time 2.133 seconds >>perft 5=1260017972 time 119.571 seconds >> >>yace agrees about the numbers and I know of no bugs of the program(I already >>corrected the known problem of the enpassant that is not relevant here) >> >>I still suspect that there are unknown bugs because the program is very >>complicated. > >Hi, Uri! I am not sure what your program have said. Is the program, that you >used, found mate in 10 moves, or at other depth, maybe even much shorter that >ten? No My program is not designed to find mates but only to generate moves and I test it by giving it positions and ask it to calculate the perft function. In the initial position perft 1=20 because white has exactly 20 legal moves perft 2=400 because there are exactly 400 legal games of 2 plies. I believe that my program is better than all the free program that I know. I know that chest also can calculate perft and it is probably the best free program in doing it. Note that chest is not optimized for the task and usually programs have this function only to check for bugs and they did not optimize this function for speed. I use it to check for bugs but I also try to optimize it for speed but I am sure that it is possible to do it clearly faster by assembler when I use only C with microsoft visual c++6 compiler. 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.