Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: crafty 16.3 - something broken in K and P endgames?

Author: Robert Hyatt

Date: 14:58:27 02/05/99

Go up one level in this thread


On February 05, 1999 at 00:13:59, James B. Shearer wrote:

>On January 31, 1999 at 10:43:10, Robert Hyatt wrote:
>
>>On January 30, 1999 at 23:54:15, James B. Shearer wrote:
>>
>>>       I was playing Fitter on ICC and was surprised when it let me escape with
>>>a draw in a K and P endgame.  I checked the key position afterwards with crafty
>>>16.3 (on a 233 k6 PC).  Crafty 16.3 takes 2 hours to find the winning move, h4,
>>>at ply 16 (see log below).  On the other hand crafty 15.15 finds h4 in 8 seconds
>>>at ply 10.  Fritz 4 also finds h4 in a few seconds.  Looks like you may have
>>>broken something.
>>>                                James B. Shearer
>>
>>
>>you must have something broken on your machine, first.  I ran this on my
>>PII/300 notebook, and get to ply 16 in a minute, not hours.  Is your machine
>>swapping/paging a lot?  It takes me to depth=16 to fail high on h4, takes 2:00
>>on my notebook, and 25 seconds on my quad xeon...  I couldn't begin to explain
>>why it would take hours to get to ply=16 there...  I am at ply=17 failing high
>>again after only 3 minutes on my notebook.
>
>         I think my machine is ok.  I ran the position again using the annotate
>command for 7200 seconds (2 hours) and the node rate stayed at 60000 same as for
>a 60 second search so I don't think swapping/paging is the problem.  See log
>below.
>                               James B. Shearer
>hash table memory = 12M bytes.
>pawn hash table memory = 2M bytes.
>EGTB access enabled
>using tbpath=./TB
>0 piece tablebase files found
>pondering disabled.
>book learning disabled
>result learning disabled
>position learning disabled
>noise level set to 0.
>
>
>Crafty v16.3
>
>White(1): annotate fitkp.pgn b 40-40 -1 7200
>
><snip moves>
>
>Black(40): g5
>
>              Searching all legal moves.----------------------------------
>end-game phase
>              clearing hash tables
>              time surplus   0.00  time limit 120:00 (120:00)
>              depth   time  score   variation (1)
>                1     0.00   1.94   40. ... g5
>                1     0.01   1.96   40. ... Kg5
>                1->   0.02   1.96   40. ... Kg5
>                2     0.03     --   40. ... Kg5
>                2     0.03   0.00   40. ... Kg5
>                2     0.03     ++   40. ... h4!!
>                2     0.04   1.57   40. ... h4 41. gxh4 Kf4
>                2     0.05   1.89   40. ... g5 41. Ke3
>                2->   0.06   1.89   40. ... g5 41. Ke3
>                3     0.06   1.69   40. ... g5 41. Ke3 Kg6
>                3->   0.07   1.69   40. ... g5 41. Ke3 Kg6
>                4     0.08   1.81   40. ... g5 41. Ke3 g4 42. h4
>                4->   0.10   1.81   40. ... g5 41. Ke3 g4 42. h4
>                5     0.13   1.87   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7
>                5->   0.15   1.87   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7
>                6     0.18   1.61   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7 44. Ke3
>                6->   0.20   1.61   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7 44. Ke3
>                7     0.32   1.77   40. ... g5 41. Ke3 g4 42. h4 Kg6 43.
>                                    Kf4 Kf7
>                7->   0.35   1.77   40. ... g5 41. Ke3 g4 42. h4 Kg6 43.
>                                    Kf4 Kf7
>                8     0.73   1.81   40. ... g5 41. Ke3 Kg6 42. Ke2 g4 43.
>                                    h4 Kf5 44. Ke3
>                8->   0.77   1.81   40. ... g5 41. Ke3 Kg6 42. Ke2 g4 43.
>                                    h4 Kf5 44. Ke3
>                9     1.48   1.71   40. ... g5 41. Ke3 Kg6 42. Ke2 g4 43.
>                                    h4 Kf7 44. Ke3 Ke7
>                9->   1.53   1.71   40. ... g5 41. Ke3 Kg6 42. Ke2 g4 43.
>                                    h4 Kf7 44. Ke3 Ke7
>               10     2.49   1.71   40. ... g5 41. Ke3 Kg6 42. Kf2 g4 43.
>                                    h4 Kf7 44. Ke3 Ke7 45. b4 <HT>
>               10->   2.57   1.71   40. ... g5 41. Ke3 Kg6 42. Kf2 g4 43.
>                                    h4 Kf7 44. Ke3 Ke7 45. b4 <HT>
>               11     5.69   1.71   40. ... g5 41. Ke3 Kg6 42. Kf2 g4 43.
>                                    h4 Kf7 44. Ke2 Ke7 45. Ke3 Kd7
>               11->   5.81   1.71   40. ... g5 41. Ke3 Kg6 42. Kf2 g4 43.
>                                    h4 Kf7 44. Ke2 Ke7 45. Ke3 Kd7
>               12     9.15   1.71   40. ... g5 41. Ke3 Kg6 42. Kf2 g4 43.
>                                    h4 Kf7 44. Ke2 Ke7 45. Ke3 Kd7 46.
>                                    b4
>               12->   9.89   1.71   40. ... g5 41. Ke3 Kg6 42. Kf2 g4 43.
>                                    h4 Kf7 44. Ke2 Ke7 45. Ke3 Kd7 46.
>                                    b4
>               13    26.26   1.71   40. ... g5 41. Ke3 Kg6 42. Kf2 g4 43.
>                                    h4 Kf7 44. Ke2 Ke7 45. Kf2 Kd7 46.
>                                    Ke3 Kc6
>               13->  28.44   1.71   40. ... g5 41. Ke3 Kg6 42. Kf2 g4 43.
>                                    h4 Kf7 44. Ke2 Ke7 45. Kf2 Kd7 46.
>                                    Ke3 Kc6
>               14     3:10   1.72   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7 44. Kf2 Ke7 45. Ke2 Kd7 46.
>                                    Kf2 Kd8 47. Ke3 Ke7 48. c4
>               14->   3:14   1.72   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7 44. Kf2 Ke7 45. Ke2 Kd7 46.
>                                    Kf2 Kd8 47. Ke3 Ke7 48. c4
>               15     4:57   1.71   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7 44. Kf2 Ke7 45. Ke2 Kd7 46.
>                                    Kf2 Kd8 47. Ke3 Kc7 48. Kf2 Kc6 49.
>                                    Ke3
>               15->   5:14   1.71   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7 44. Kf2 Ke7 45. Ke2 Kd7 46.
>                                    Kf2 Kd8 47. Ke3 Kc7 48. Kf2 Kc6 49.
>                                    Ke3
>               16    11:46   1.71   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7 44. Kf2 Ke7 45. Ke2 Kd7 46.
>                                    Kf2 Kd8 47. Ke3 Kc7 48. Ke2 Kc6 <HT>
>               16->  12:39   1.71   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7 44. Kf2 Ke7 45. Ke2 Kd7 46.
>                                    Kf2 Kd8 47. Ke3 Kc7 48. Ke2 Kc6 <HT>
>               17    53:37   1.71   40. ... g5 41. Ke3 h4 42. g4+ Kg6 43.
>                                    Ke2 Kf7 44. Kf2 Ke7 45. Ke2 Kd7 46.
>                                    Kf2 Kd8 47. Ke3 Kc7 48. Ke2 Kc6 49.
>                                    Ke3 Kd7
>              time=120:00  cpu=99%  mat=0  n=436256725  fh=4%  nps=60589
>              ext-> checks=49824007 recaps=1196664 pawns=18327198 1rep=14431267
>              predicted=0  nodes=436256725  evals=82091710
>              endgame tablebase-> probes done=0  successful=0  maxd=32
>              hashing-> trans/ref=32%  pawn=1%  used=w99% b99%
>White(41): quit
>execution complete.


this says something is wrong.  pawn=1% means you are running about 1/2 the
speed you normally should.  Check the size of hashp as it is either bogus or
something else is broken.

fh=4% is impossible.  That means that in only 4% of the nodes where the search
fails high it fails high on the first move.  The lowest value I have ever seen
here is 88%, and it is normally above 90%.  Is this a stock crafty, or one you
compiled yourself?  IE it is possible that the compiler is breaking something.



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.