Thinker 4CPU - Fritz 11

[Werner]

I forgot to ask you werner if with the new version of Thinker you had any problems with adapter.eng with the Single CPU version? Or you tried Wb2UCI so far?

So far these bugs are with MP version of Thinker?

- I only played with 4CPU mp version
- last bug I had inside Chessbase with wb2uci adapter
- last bug could be repeated inside Arena with 4CPU Version
"edited: just able to see the same move inside Arena:
FEN: 2r2k2/2q2p1r/pp1bp1p1/n3N3/3P1PQ1/P1P4P/6RP/2B2RK1 w - - 4 26

(Passive)Thinker_64-bit-smp4:
Th8-h7
13 00:20 0 0 +0,77 h8h7 c1b2 a5c4"
best

[DomLeste]

Ok i be busy with Rybka myself when u guys start with single CPU version of Thinker 5.2E tell us which adapter you use It could be the MP only as its new who knows perhaps lance may say something.

Dom

[CThinker]

Here is my result so far:

Code: Select all

1   Fritz 11                +27/-3/=2 87.50%   28.0/32
2   Thinker 5.2E x64p 4CPU  +3/-27/=2 12.50%    4.0/32

This is a very bad result - behind 1CPU of 5.2D.
Are there any Problems?

Here is 1 game:

[Event "SPEEDSTER, 10'/40+10'/40+10'/40"]
[Site "SPEEDSTER"]
[Date "2008.08.10"]
[Round "22"]
[White "Fritz 11"]
[Black "Thinker 5.2E x64p 4CPU"]
[Result "1-0"]
[ECO "B06"]
[Annotator "-0.12;-0.17"]
[PlyCount "61"]
[TimeControl "40/600:40/600:40/600"]

{Intel(R) Core(TM)2 Quad CPU @ 2.66GHz 3504 MHz W=17.0 ply; 2.803kN/
s; HS-8moves.ctg B=15.9 ply; 0kN/s; HS-8moves.ctg} 1. e4 {B/0 0} g6 {B/0 0} 2.
d4 {B/0 0} Bg7 {B/0 0} 3. f4 {B/0 0} d5 {B/0 0} 4. e5 {B/0 0} Nh6 {B/0 0} 5.
Nf3 {Beide letzter Buchzug B/0 0} O-O {-0.17/17 14} 6. Bd3 {(Sc3) -0.12/17 16}
c5 {(Sc6) -0.09/16 14} 7. dxc5 {-0.02/17 19} Qc7 {(Da5+) -0.20/16 18} 8. Be3 {
0.77/17 21} e6 {(Le6) 0.20/16 21} 9. Nc3 {(Dd2) 1.16/16 11} Ng4 {
(Ld7) 0.48/16 11} 10. Nb5 {(Dd2) 1.42/17 12} Qe7 {0.29/16 17} 11. Qe2 {
(Ld4) 1.36/17 16} Nxe3 {(f6) 0.32/16 17} 12. Qxe3 {1.38/16 3} b6 {0.46/16 15}
13. Nd6 {(cxb6) 1.16/17 19} Nd7 {(bxc5) 0.93/15 13} 14. c6 {(Sxc8) 2.07/17 11}
Nc5 {1.46/17 54} 15. O-O {(Sxc8) 2.10/17 13} f6 {(Td8) 1.18/15 21} 16. Nd4 {
(Sxc8) 2.56/15 13} a5 {(g5) 0.96/14 11} 17. Rad1 {(Sxc8) 2.56/14 16} Qd8 {
(Td8) 1.56/14 16} 18. Qg3 {(Sxc8) 2.65/14 18} Qc7 {(De7) 1.67/14 16} 19. h4 {
(Sxc8) 2.77/14 14} Qe7 {(Kh8) 1.29/16 34} 20. h5 {(Sxc8) 2.78/16 12} Nxd3 {
(gxh5) 1.76/16 61} 21. cxd3 {(Txd3) 2.98/16 9} gxh5 {1.74/14 13} 22. Rc1 {
(Dh4) 3.08/17 29} Kh8 {(b5) 1.35/16 9} 23. c7 {(Sxc8) 3.48/16 7} h6 {
(La6) 2.43/16 37} 24. Qh4 {(Sxc8) 4.42/15 11} Qd7 {(b5) 3.35/16 73} 25. Qxh5 {
4.21/16 9} a4 {(De7) 2.56/14 9} 26. a3 {(Sxc8) 4.74/15 11} fxe5 {
(De7) 2.72/14 7} 27. fxe5 {7.87/16 4} Ba6 {(Lb7) 5.92/13 48} 28. Nf7+ {
(Txf8+) 9.34/16 7} Rxf7 {(Kg8) 20.19/14 24} 29. Rxf7 {12.08/18 3} Qe8 {
10.92/14 8} 30. Qg6 {(Tc6) 23.27/18 8} Qg8 {217.88/13 8} 31. Nxe6 {#15/18 5}
1-0

Sorry about this one. I introduced some simplification in the way that the best root move is saved in a root split situation. I thought this was a good thing, and I spent not enough time testing it. As it turns out, only the first thread is able to save the best move. When other threads find the best move, it does not get saved.

For now, please stop using the SMP version. I will fix this, and also add the ability to configure the number of threads.

Again, my apologies.

[Werner]

[quote="CThinkerSorry about this one. I introduced some simplification in the way that the best root move is saved in a root split situation. I thought this was a good thing, and I spent not enough time testing it. As it turns out, only the first thread is able to save the best move. When other threads find the best move, it does not get saved.

For now, please stop using the SMP version. I will fix this, and also add the ability to configure the number of threads.

Again, my apologies.[/quote]

Hi Lance,
there is really no need for apologies
Your engine is free, I like to test engines and I am sure you find a good solution. Then we shall continue testing.

thanks