Author: Eduard Nemeth
Date: 02:13:44 03/20/05
Mysteriöse Analyse - und Zug von Shredder 9 UCI mit 128 MB Hashtables!
Die Stellung und Partie fand ich in dieser message:
http://f23.parsimony.net/forum50826/messages/122747.htm
[Event "Me1.2"]
[Site "MyTown"]
[Date "2004.12.31"]
[Round "1.1"]
[White "CM10000_Me1.2"]
[Black "Shredder 9"]
[Result "1-0"]
[ECO "E04"]
[PlyCount "59"]
[EventDate "2004.12.31"]
[EventType "tourn"]
[TimeControl "1800+5"]
1. d4 {0} Nf6 {0} 2. c4 {0} e6 {0} 3. Nf3 {0} d5 {0} 4. g3 {0} dxc4 {0} 5. Bg2
{0} Nc6 {0} 6. Qa4 {0} Bb4+ {0} 7. Bd2 {0} Nd5 {0} 8. Qb5 {0} O-O {0} 9. Qxc4 {
0} Nb6 {0} 10. Qd3 {0} Be7 {0} 11. O-O {-0.03/11 39} Qd5 {(e5) 0.30/15 136} (
11... Qd5 {90}) 12. Ng5 {(Sc3) 3.04/12 37} Qf5 {3.55/17 151} 13. Be4 {
3.00/12 44} Qf6 {3.55/18 125} 14. Bxc6 {(Lxh7+) 3.00/10 39} Qh6 {3.09/16 64}
15. Nf3 {3.25/12 47} Qh3 {3.17/16 139} 16. Bb5 {3.49/12 55} a6 {3.20/16 66} 17.
Bc4 {3.43/11 21} Bd7 {3.24/16 63} 18. Ne5 {3.74/11 49} Rad8 {3.36/15 67} 19.
Bb3 {3.88/11 61} Bf6 {3.44/15 58} 20. Qe4 {4.05/11 48} Ba4 {3.58/15 54} 21.
Qxb7 {(Lxa4) 4.18/11 55} Bxb3 {(Txd4) 3.68/16 54} 22. axb3 {4.34/11 26} Rxd4 {
3.84/17 63} 23. Nd3 {(Sf3) 4.56/11 81} Rd7 {(Sd5) 3.94/14 51} 24. Rxa6 {
(Sc3) 5.04/11 65} Rfd8 {(Sd5) 4.21/15 43} 25. Nc3 {5.17/10 27} Qh5 {4.64/14 66}
26. Qc6 {(Lf4) 5.26/11 47} Bd4 {4.52/14 41} 27. Rfa1 {5.52/10 31} Bxc3 {
4.55/14 31} 28. Nf4 {6.10/11 28} Qf5 {(Dg5) 4.91/15 45} 29. Bxc3 {6.32/11 39}
Qg5 {(Dc2) 5.73/16 181} 30. Ra7 {6.72/11 42} 1-0
Ich konnte den Zug reproduzieren!
Step1: Ich löschte die Pl2 - Lerndatei im Ordner von Shredder 9 UCI
Step2: Ich startete die GUI (ohne dass Shredder 9 rechnete) und stellte 128 MB
Hash ein
Step3: Ich startete die folgende Analyse unten
[D]r1bq1rk1/ppp1bppp/1nn1p3/8/3P4/3Q1NP1/PP1BPPBP/RN3RK1 b - -
Shredder 9 played Qd5???
Engine: Shredder 9 UCI (128 MB)
von Stefan Meyer-Kahlen
9/21 0:01 -0.24 1...Lf6 2.Lc3 Sd5 3.e3 Dd6 4.Sbd2 Sxc3
5.bxc3 Ld7 6.Tab1 (118.401) 99
10/25 0:02 -0.18 1...Lf6 2.Lc3 Sa4 3.Sbd2 Sxc3 4.bxc3 e5
5.e3 Le6 6.Se4 exd4 (274.435) 101
11/26 0:05 -0.42 1...Lf6 2.Lc3 Sd5 3.Sbd2 (518.340) 101
11/26 0:06 -0.41++ 1...f6 (623.493) 102
11/26 0:09 -0.31 1...f6 2.e3 Dd6 3.Sc3 e5 4.Se4 (953.492) 102
12/31 0:29 -0.32 1...f6 2.Tc1 Ld7 3.e3 De8 4.Sc3 Dh5
5.Se4 Sd5 6.Sc5 Scb4 7.Lxb4 (3.039.157) 101
12/31 0:31 -0.31++ 1...Lf6 (3.197.933) 101
12/31 0:32 -0.31 1...Lf6 2.Lc3 Sd5 3.Sbd2 Te8 4.Se4 (3.267.550) 101
13/25 0:44 -0.33 1...Lf6 2.Lc3 Dd5 3.Sbd2 Dh5 4.e3 Sd5 (4.609.524) 102
13/29 0:59 -0.32++ 1...Te8 (6.081.865) 102
13/29 1:09 -0.32 1...Te8 2.Tc1 (7.177.050) 102
14/31 2:13 -0.26 1...Te8 2.Lc3 Dd5 (13.686.606) 102
15/32 3:20 -0.35 1...Te8 2.Sc3 Lf6 3.Tad1 e5 4.dxe5 Dxd3
5.exd3 Sxe5 6.Sxe5 Lxe5 7.Tde1 Te6
8.Le4 Te7 9.Sd1 (20.475.909) 102
15/32 4:09 -0.34++ 1...f6 (25.262.993) 101
15/32 4:20 -0.34 1...f6 2.Tc1 Ld7 3.e4 e5 4.Db3+ Tf7
5.d5 (26.373.822) 101
15/32 4:35 -0.33++ 1...Lf6 (27.937.323) 101
15/32 4:41 -0.32 1...Lf6 2.Lc3 Sd5 3.Sbd2 Sxc3 4.bxc3 e5
5.d5 Se7 6.e4 Lg4 7.Sc4 Sg6 8.Tab1 (28.570.069) 101
15/32 5:17 -0.31++ 1...g5 (32.245.586) 101
15/32 5:41 -0.31 1...g5 2.e3 (34.856.317) 102
15/32 6:00 -0.30++ 1...Dd5 (36.725.389) 101
15/35 6:47 -0.30 1...Dd5 2.Sc3 Dh5 3.Lf4 (41.186.199) 101
16/30 9:15 -0.55-- 1...Dd5 2.Sc3 (55.949.068) 100
16/30 9:24 -1.05-- 1...Dd5 2.Sc3 (56.827.778) 100
16/50 15:53 -3.65 1...Dd5 2.Sg5 Df5 3.Le4 Df6 4.Lxh7+ Kh8
5.g4 g6 6.Dh3 Dg7 7.Lxg6+ Kg8 8.Lh7+ Kh8
9.Le4+ Kg8 10.Sf3 Lf6 11.Lh6 Dh8
12.Dh5 Sxd4 13.Sxd4 Lxd4 14.Lxf8 Dxh5 (100.261.398) 105
16/50 15:56 -3.64++ 1...Te8 (100.652.604) 105
16/50 16:03 -3.29++ 1...Te8 (101.465.851) 105
16/50 17:29 -0.38 1...Te8 2.Sc3 Lf6 3.Tad1 e5 4.dxe5 Dxd3
5.exd3 Sxe5 6.Sxe5 Lxe5 7.Tde1 Te6
8.Le3 c6 9.Le4 (109.184.012) 104
16/50 18:38 -0.37++ 1...f6 (115.070.392) 102
Auf meinem natürlich sehr langsamen Rechner (600 MHz) griff S9 zu Qd5?? Nach
exakt 6 Min. und hielt den Zug bis fast 16 Min!
Da muss was falsch laufen bei Shredder 9, denn bei anderen Hashtabellen-Größen
zeite suich dieses Phänomen nicht!
Eine mögliche Erklärung dazu hat vorhin Dr. Christian Donninger
(Hydra-Programmierer) gegeben, in dieser message:
http://f23.parsimony.net/forum50826/messages/122798.htm
Was meinen die Experten hier dazu?
Eduard
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.