Author: leonid
Date: 08:14:00 10/09/01
Go up one level in this thread
On October 09, 2001 at 11:04:26, Heiner Marxen wrote: >On October 08, 2001 at 18:43:14, leonid wrote: > >>On October 07, 2001 at 13:46:06, Heiner Marxen wrote: >> >>>On September 30, 2001 at 05:43:43, Tim Foden wrote: >>> >>>>On September 29, 2001 at 07:33:59, leonid wrote: >>>> >>>>>Hello! >>>>> >>>>>This mate is probably very easy. There are good chance that you will solve it >>>>>even by brute force wihtout too much waiting. >>>> >>>>As other post pointed out, I'm afraid this one isn't as easy as you thought :) >>> >>>Yes. Chest agrees with you :-) >>> >>> >>>>>[D]RQ2rqbb/4p2k/1KQQqQ1n/1R2qqpq/1P3P2/1BBPNN2/nq6/rq4QQ w - - >>>>> >>>>>Please indicate your result. >>>> >>>>GLC215-pr9, 920MHz Duron, 96MB hash claims Nxg5+ is a mate in 11, after 10m 3s. >>>> >>>>Cheers, Tim. >>> >>>Chest on K7/600 with 350 MB hash needs 3.6 days (!) to be sure about it. >>>1.Nxg5+ is the only key move for mate in 11. Interestingly, Chest gives >>>the exact same PV as GLC below. >>>To prove that there is no mate in 10 Chest needs more than 12 hours, already. >>> >>>I wonder how Leonid did this one with "brute force search". Leonid? >> >> >>Hi, Heiner! >> >>I went on this position only 10 moves deep. Since mine found mate in 11, I was >>sure that it is mate in 11. Probably will put this position by brute force at >>night 11 moves deep. >> >>I expected that some programs with more hash and quiker computer will solve more >>rapidly that mine. This situation I was able to see repeatedly with your >>program. Sometime I feel that I must to come back to my code and do something. >> >>Brute force search: >> >>Depth (moves) Time Branching factor nodes/second >> >>4 moves 0.93 sec 77k (killo nodes/one sec) >> 4.0 >>5 moves 3.73 sec 72k >> 3.94 >>6 moves 14.72 sec 59k >> 4.29 >>7 moves 63 sec 65k >> 4.06 >>8 moves 4 min 16 sec 74k >> 5.8 >>9 moves 24 min 54 sec 78k >> 5.5 >>10 moves 2 h 18 min 27 sec 88k >> >>Celeron 600Mhz. >> >>Enjoyed very much your program!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! >> >>Really very, very impressive! > >Thanks for your kind words! > >The above times for your program clearly indicate that Chest can be >improved. This time you are 5 times as fast as Chest! >Will have to study that problem later. Thanks, but this is occasional. In the most cases your brute force is much speedier that mine. It is, probably, me who must seriously start looking into his code. For now I have all kind of excuses for not doing this. I put for night this position on 11 moves. Time was 6 hours and 51 min. I still wait for your graphics version. Will be great for everybody when you will finish it. All the best in your work! Cheers, Leonid. >Cheers, >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.