Author: Uri Blass
Date: 13:48:05 12/06/01
Go up one level in this thread
On December 06, 2001 at 15:41:03, Severi Salminen wrote: >>The point is that I do not have to do the extra work to calculate perft 6 nad >>calculating perft is a well defined problem that is different from playing >>chess. >> >>There is a point in comparing the time that my program needs to do a job with >>the time that other programs need to do the same task. > >True, but I consider perft as a task where engine updates the board >representation structure (or array) to all legal positions from a certain >position. So I think that one really should visit also the leaf nodes, not just >generate the moves (even if they are legal). I can do perft 5 in a second if I >do all but the "frontier-moves". But I don't think it is right number - Crafty >also makes/unmakes ALL the moves. And this means updating the structures. And if >we don't do the same we shouldn't compare elapsed time also. Crafty does not do the same job that I do when it makes move so comparison when I have to make the last ply is not fair. I can also add that I use only C and no assembler tricks so I am sure that it is possible to improve the speed of my program but the main improvenet can come out of other possible ideas(I can save most of the nodes that I generate by trying to see if I can change the order of moves after some plies. 1.e4 e5 2.d4 d5 1.e4 d5 2.d4 e5 1.d4 e5 2.e4 d5 1.d4 d5 2.e4 e5 are leading to exactly the same position and it is possible to use this fact to save nodes. I will not complain if another program is using it and is faster than mine. I think that the only fair comparison with other program is time. It is not my fault that other programs are not optimized to calculate perft. I believe that there is no free program that is faster than mine in the task that is well defined. > >>I did not decide about a name for my program. >>I may also try to test it for more positions to see if I have more bugs. >> >>It is very complicated so I am afraid from bugs. >>I found bugs even after finding that it calculates perft 8 correctly. > >That is very likely (not that it is likely that YOU make bugs, but that you >don't maybe notice them in 8 ply search) as there are no promotions or queenside >castlings in that depth (from initial position, of course). my bug was not about it. I checked also other position when all castling and promotion were possible even before calculating perft 8 from the initial position and there was no problem in calculating perft 5. 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.