Author: Pete Galati
Date: 16:07:22 02/08/01
Go up one level in this thread
On February 08, 2001 at 17:08:36, leonid wrote: >On February 08, 2001 at 11:42:40, Pete Galati wrote: > >>On February 08, 2001 at 11:31:35, leonid wrote: >> >>>On February 08, 2001 at 10:45:07, Heiner Marxen wrote: >>> >>>>On February 08, 2001 at 10:09:29, leonid wrote: >>>> >>>>>On February 08, 2001 at 09:49:16, Pete Galati wrote: >>>>> >>>>>>On February 08, 2001 at 08:21:44, leonid wrote: >>>>>> >>>>>>>Hi! >>>>>>> >>>>>>>If you like to solve easy (not deep) forced mate position, here is the one. >>>>>>> >>>>>>>[D]Nb1rkrNb/1n1pqp1b/n2qqq1n/1NNNNNNn/2BNNB2/1R2Q1R1/8/2K5 w - - >>>>>>> >>>>>>>Please, indicate your result. >>>>>>> >>>>>>>Thanks, >>>>>>>Leonid. >>>>>> >>>>>>Crafty found the mate in 8 a little bit earlier but it had it's hashtables all >>>>>>filled up, and I can't operate it at up to 96mb hashtables on my computer, so >>>>>>I'm stuck with 48 at the moment. I never get around to mentioning, it's a >>>>>>700mhz Celeron. >>>>>> >>>>>>Pete >>>>>> >>>>>>5 piece tablebase files found >>>>>>1836kb of RAM used for TB indices and decompression tables >>>>>>EGTB cache memory = 3M bytes. >>>>>>hash table memory = 48M bytes. >>>>>>pawn hash table memory = 5M bytes. >>>>> >>>>>Simply can't detain my curiosity, what is the "pawn hash table"? Special hash >>>>>table just for pawns? >>>> >>>>Yes, for that part of the eval which is based only on the location of >>>>all the pawns (pawn structure). >>>> >>>> >>>>>Thanks for detailed decription! Sometime, some very interesting moments come to >>>>>light. >>>>> >>>>>My program found at the same depth mate but I not looked by brute force. It >>>>>could be that somebody will come with shorter mate. >>>>> >>>>>Leonid. >>>> >>>>I doubt it. Chest says "no mate in 7" after 68.4 minutes. >>>>You all found the shortest mate: Congratulations! >>> >>>Thanks, but I doubt very much that my program will search 7 moves in one hour. >>>It had awful branching factor and in 5 moves already went to 10 min. Selective >>>search make me feel this by solving position by easiest level (8 moves deep)in >>>18 sec. Usually it should be only split of one second for 8 moves, for that >>>search level. >>> >>>Brute force search: >>> >>>3 moves - 0.38 sec >>>4 moves - 22.3 sec >>>5 moves - 9 min 52 sec >>> >>>Leonid. >> >>Here's a paragraph from Bob's Crafty doc file related to the pawn hashtables: >> >>"23. hash=x and hashp=x These commands are used to adjust >>the size of the hash tables in Crafty. hash modifies the >>size of the transposition/refutation table, while hashp mod- >>ifies the size of the pawn structure/king safety hash table. >>The sizes may be entered as one of the following two types >>of values: nnnK where nnn is an integer indicating how many >>Kbytes Crafty should use for this hash table; nnnM where nnn >>is an integer indicating how many Mbytes Crafty should use." >> >>Pete > >Thanks, Pete! > >Now I see that Crafty code should be completely different from everything that I >have in my program. Difference is everywhere! The first impression I had when >I looked into Kerrigan description and he said me that its code is "like usual". > >Leonid. I'm not sure what Tom would have meant by "like usual". I suppose if you look at Crafty's individual functions, you might not see too many UNusual things, but when you look at Crafty as a whole, I would call it a text book in Chess and C programming. It is extensive, but it's not too difficult to find your way around with something like grep. I haven't looked at the Crafty code for a while, but it's well worth a tour if you have the time to do it. Pete
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.