Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: How to test hashtable implementation?

Author: Uri Blass

Date: 10:31:59 07/20/01

Go up one level in this thread


On July 20, 2001 at 12:24:06, Miguel A. Ballicora wrote:

>On July 20, 2001 at 11:39:06, Adam Oellermann wrote:
>
>>
>>>Do you mean the Fine #70 endgame position?
>>>Here is it:
>>>
>>[D]8/k7/3p4/p2P1p2/P2P1P2/8/8/K7 w - - 0 1
>>>
>>>solution starts with 1.Kb1 Kb7 2.Kc1
>>>
>>>Rafael B. Andrist
>>
>>Yup, that's the one. Any info on what depth is required to solve it, and how
>>long it might take?
>
>Anything from 18 to 26 plies is ok (B. Hyatt posted this) . It varies according
>to the program, and it should take few seconds, may be less than a second in
>most computers to go for Kb1. In a PIII 700mhz 16 Mb Hash table mine takes 0.2 s
>to reach ply 21 which is the first one that chooses Kb1. Later, I had a big
>increase in score at ply 26.
>My program has almost no knowledge relevant to this endgame (just centralization
>of the king and material since there are no passed pawns) so I guess it would be
>interesting for you to compare.
>
>I posted long time ago (January this year I think) some experiments I did
>with this position and some combinations of hashtable schemes and evaluation
>functions. I believe that you might find it useful, sometimes you might find
>the solution is less plies but in longer time!
>
>This is the output (Gaviota 0.2.2)
>
>  Ply    seconds  score  PV
>  18:      0.1    +0.62  Ka1-b2   Ka7-b6   Kb2-c3   Kb6-c7   Kc3-c4   Kc7-b6
>                         Kc4-d3   Kb6-c7   Kd3-c3   Kc7-b7   Kc3-d2   Kb7-c8
>                         Kd2-c2   Kc8-d7   Kc2-b2   Kd7-c7   Kb2-b3   Kc7-b6
>  19       0.1    +0.62  Ka1-b2   Ka7-b8   Kb2-c3   Kb8-b7   Kc3-d2
>  19:      0.1    +0.62  Ka1-b2   Ka7-b8   Kb2-c3   Kb8-b7   Kc3-d2
>  20       0.1    +0.62  Ka1-b2   Ka7-a8   Kb2-c3   Ka8-b7   Kc3-d2
>  20:      0.1    +0.62  Ka1-b2   Ka7-a8   Kb2-c3   Ka8-b7   Kc3-d2
>  21       0.1    +0.62  Ka1-b2   Ka7-a8   Kb2-c3   Ka8-b7   Kc3-d2   Kb7-c8
>                         Kd2-c2   Kc8-b8   Kc2-d3   Kb8-c7   Kd3-e3   Kc7-d7
>                         Ke3-e2   Kd7-d8   Ke2-f3   Kd8-e7   Kf3-g3   Ke7-f6
>                         Kg3-h4   Kf6-g6   Kh4-g3
>  21       0.2    +0.68  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-e7   Kc2-d3   Ke7-f7   Kd3-c4   Kf7-g6
>                         Kc4-b5   Kg6-h5   Kb5-c6   Kh5-g4   Kc6xd6   Kg4xf4
>                         Kd6-c5   Kf4-e4    d5-d6
>  21:      0.2    +0.68  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-e7   Kc2-d3   Ke7-f7   Kd3-c4   Kf7-g6
>                         Kc4-b5   Kg6-h5   Kb5-c6   Kh5-g4   Kc6xd6   Kg4xf4
>                         Kd6-c5   Kf4-e4    d5-d6
>  22       0.2    +0.63  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-e7   Kc2-d3   Ke7-f7   Kd3-c4   Kf7-g6
>                         Kc4-b5   Kg6-h5   Kb5-c6   Kh5-g4   Kc6xd6   Kg4xf4
>                         Kd6-c5   Kf4-e4    d5-d6    f5-f4
>  22:      0.2    +0.63  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-e7   Kc2-d3   Ke7-f7   Kd3-c4   Kf7-g6
>                         Kc4-b5   Kg6-h5   Kb5-c6   Kh5-g4   Kc6xd6   Kg4xf4
>                         Kd6-c5   Kf4-e4    d5-d6    f5-f4
>  23       0.2    +0.74  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-e7   Kc2-d3   Ke7-f7   Kd3-c4   Kf7-g6
>                         Kc4-b5   Kg6-h5   Kb5-c6   Kh5-g4   Kc6xd6   Kg4xf4
>                         Kd6-c5   Kf4-e4    d5-d6    f5-f4    d6-d7
>  23:      0.2    +0.74  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-e7   Kc2-d3   Ke7-f7   Kd3-c4   Kf7-g6
>                         Kc4-b5   Kg6-h5   Kb5-c6   Kh5-g4   Kc6xd6   Kg4xf4
>                         Kd6-c5   Kf4-e4    d5-d6    f5-f4    d6-d7
>  24       1.2      :-)  Ka1-b1
>  24       2.9    +0.73  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-c8   Kc2-d2   Kc8-d7   Kd2-c3   Kd7-c7
>                         Kc3-d3   Kc7-b7   Kd3-d2   Kb7-c8   Kd2-c2   Kc8-b8
>                         Kc2-b3   Kb8-c7   Kb3-c3   Kc7-b7   Kc3-d3   Kb7-c7
>  24:      2.9    +0.73  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-c8   Kc2-d2   Kc8-d7   Kd2-c3   Kd7-c7
>                         Kc3-d3   Kc7-b7   Kd3-d2   Kb7-c8   Kd2-c2   Kc8-b8
>                         Kc2-b3   Kb8-c7   Kb3-c3   Kc7-b7   Kc3-d3   Kb7-c7
>  25       2.9      :-)  Ka1-b1
>  25      23.4    +0.73  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-c8   Kc2-d2   Kc8-d7   Kd2-c3   Kd7-c7
>                         Kc3-d3   Kc7-b7   Kd3-e3   Kb7-c8   Ke3-f3   Kc8-d7
>                         Kf3-e2   Kd7-d8   Ke2-d3   Kd8-c7   Kd3-c4   Kc7-b6
>                         Kc4-d3
>  25:     23.4    +0.73  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-c8   Kc2-d2   Kc8-d7   Kd2-c3   Kd7-c7
>                         Kc3-d3   Kc7-b7   Kd3-e3   Kb7-c8   Ke3-f3   Kc8-d7
>                         Kf3-e2   Kd7-d8   Ke2-d3   Kd8-c7   Kd3-c4   Kc7-b6
>                         Kc4-d3
>  26      23.4      :-)  Ka1-b1
>  26     144.3    +1.40  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-c8   Kc2-d2   Kc8-d7   Kd2-c3   Kd7-c7
>                         Kc3-d3   Kc7-b7   Kd3-e3   Kb7-c8   Ke3-f3   Kc8-d7
>                         Kf3-g3   Kd7-e7   Kg3-h4   Ke7-f6   Kh4-h5   Kf6-e7
>                         Kh5-g6   Ke7-d7   Kg6xf5
>  26:    144.3    +1.40  Ka1-b1   Ka7-b7   Kb1-c1   Kb7-c7   Kc1-d1   Kc7-d7
>                         Kd1-c2   Kd7-c8   Kc2-d2   Kc8-d7   Kd2-c3   Kd7-c7
>
>Regards,
>Miguel

For comparison here is the output of Junior7 on p200

It also has no problem to find Kb1 but it does not count in plies.

New position
8/k7/3p4/p2P1p2/P2P1P2/8/8/K7 w - - 0 1

Analysis by Junior 7:

1.Kb2 Kb6
  ²  (0.44)   Depth: 3   00:00:00
1.Kb2 Kb6 2.Kc3
  ²  (0.45)   Depth: 6   00:00:00
1.Kb2 Kb6 2.Kc3 Kc7 3.Kc4
  ²  (0.51)   Depth: 9   00:00:00
1.Kb2 Kb6 2.Kc3 Kc7 3.Kc4 Kb6
  ²  (0.62)   Depth: 12   00:00:00
1.Kb2 Kb6 2.Kb3 Kc7 3.Kc3 Kb6 4.Kc4
  ²  (0.62)   Depth: 15   00:00:00  2kN
1.Kb2 Kb6 2.Kb3 Kc7 3.Kc4 Kb6 4.Kd3 Ka7
  ²  (0.62)   Depth: 18   00:00:00  4kN
1.Kb2 Kb6 2.Kb3 Kc7 3.Kc4 Kb6 4.Kd3 Kc7
  ²  (0.62)   Depth: 21   00:00:00  6kN
1.Kb2 Kb6 2.Kb3 Kc7 3.Kc4 Kb6 4.Kd3 Kc7
  ²  (0.62)   Depth: 24   00:00:01  9kN
1.Kb2 Kb6 2.Kb3 Kc7 3.Kc4 Kb6 4.Kd3 Kc7
  ²  (0.62)   Depth: 27   00:00:01  14kN
1.Kb2 Kb6 2.Kb3 Kc7 3.Kc4 Kb6 4.Kd3 Kc7
  ²  (0.62)   Depth: 30   00:00:01  21kN
1.Kb2 Kb6 2.Kb3 Kc7 3.Kc4 Kb6 4.Kd3 Kc7
  ²  (0.62)   Depth: 33   00:00:01  30kN
1.Kb2 Kb6 2.Kb3 Kc7 3.Kc4 Kb6 4.Kd3 Kc7
  ²  (0.62)   Depth: 36   00:00:01  61kN
1.Kb2 Ka8 2.Kc3 Kb7 3.Kc4
  ²  (0.62)   Depth: 39   00:00:02  105kN
1.Kb1 Kb7 2.Kc1 Kc7 3.Kd1 Kd8 4.Kc2 Ke7 5.Kd3 Kf6 6.Kc4 Kg6 7.Kb5 Kh5 8.Kxa5 Kg4
9.Kb5 Kxf4 10.Kc6
  ±  (0.71)   Depth: 39   00:00:02  117kN
1.Kb1 Kb7 2.Kc1 Kc7 3.Kd1 Kd8 4.Kc2 Kc7 5.Kd3
  ±  (0.92)   Depth: 42   00:00:02  142kN
1.Kb1!
  ±  (1.22)   Depth: 45   00:00:03  188kN
1.Kb1!
  +-  (1.52)   Depth: 48   00:00:04  284kN
1.Kb1!
  +-  (1.82)   Depth: 50   00:00:07  547kN, tb=2

(blass, tel-aviv 20.07.2001)

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.