Author: Heiner Marxen
Date: 13:03:20 03/05/00
Go up one level in this thread
On March 05, 2000 at 15:36:05, leonid wrote: >On March 05, 2000 at 11:36:32, Heiner Marxen wrote: > >>On March 05, 2000 at 10:13:11, leonid wrote: >> >>>On March 04, 2000 at 18:16:54, José de Jesús García Ruvalcaba wrote: >>> >>>>On March 03, 2000 at 23:49:40, pavel wrote: >>>> >>>>>i was wondering if i need any additional hardware or software to save files in >>>>>CDs or to save files from cds to harddisk? if its hardware what should? and if >>>>>its foftware, can you pls name a good software? i am running pentium II 450mhz >>>>>compaq presario. >>>>>thanks for the help ;) >>>>> >>>>>ps, comet doesnt find a checkmate (quite easy) that almost all program find in >>>>>less then 40 sec (junior 23sec). if anyone interested i can send the game, since >>>>>i dont know how to post a diagram in this forum ;) >>>> >>>> Just paste a fen string, and add a 'D' between brackets at the left. For >>>>example: >>>> >>>>[D]6k1/6p1/8/4K1P1/8/7P/8/8 w - - 0 1 >>> >>>Tried the position for mate 9 moves deep, white to go. No mate was found. Beyond >>>this depth finding inevitable mate could be very often quit time expensive. If >>>somebody will go farther, would like to know how many moves there are before the >>>mate and how time it took. Please mention your game and your hardware. >> >>Chest 3.19 (PIII/550 500MB hash) needs 731 seconds to show that there is >>no mate in 16. Note that chest 3.19 does not access any endgame tables. >> >>There is some direct EG table access on the net. I tried >>http://chess.liveonthenet.com/chess/endings/ >>and it says: mate in 26 by Kf4. >>Since that site states to use the Thompson's CDs (which are shortest >>conversion if recall correctly), I'm not sure whether this is shortest mate. >> >>Also, this was just an example position by José de Jesús García Ruvalcaba, >>and most probably not the one, the original poster (pavel) wanted to >>talk about. >> >>Heiner >> >>>Thanks, >>>Leonid. > >Hiener you impressed me! My logic took around 1 minute to say no for 9 moves (18 >plies). It is also true that mine goes only after strict and use no memory >expensive tools, like hash tables. Anyway mine is not slow for sure. I found >this by trying the mate containing positions by thousand against the best games >in the town 4 years ago. At that time, when position contained mate, mine was >simply the speedest one after the extentive statics. Thanks! Use of hash memory can be disabled in chest. Computing the above for depth 9 needs then less than 6 seconds (PIII/550 again). :-) Well, I have a special move counting logic for the case that white has just pawns left. This does quite well, here, cutting a lot of subtrees. >All the best Heiner in your programming! I hope that you will publish somehow >you logic. Would like to see your logic expended to the complet chess game as >well. Well, it _is_ published. You can get the sources (ANSI-C) from my web page http://www.drb.insel.de/~heiner/Chess/chest.html It was never intended to be anything else but a mate solver, so I doubt it is a good start for a playing program. Feel free to try it, though :-) If by publishing you mean that I write a paper explaining it verbosely, sorry, I have no plans to do so. Some of the ideas are explained in some memo files coming with the source, some are explained in comments in the code, but some ideas are not explained there at all. Sorry. >See you next time, >Leonid. Have a nice day! Heiner
This page took 0.01 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.