Brutus v8.03 available (UCI communication bugfix)

[Stephan Vermeire]

Hi all,


Brutus v8 is available now for download at: http://www.xs4all.nl/~vermeire/brutus.html

New in version 8.03: This update contains several small UCI-communication bugfixes. Under winboard, there is no difference to 8_02.
Please let me know if you encounter MORE bugs in this version!


Best wishes,
Stephan Stephan Vermeire

[Patrick Buchmann]

Hi Stephan,

Brutus still loses on time:

[Event "Tournoi d'entrée"]
[Site "Le Fou numerique"]
[Date "2009.11.19"]
[Round "1"]
[White "Brutus 8.03"]
[Black "Abrok 5.0"]
[Result "0-1"]
[ECO "C77"]
[Opening "Spanish"]
[Time "19:52:56"]
[Variation "Centre Attack, 5...Nxd4"]
[TimeControl "300+1"]
[Termination "time forfeit"]
[PlyCount "76"]
[WhiteType "program"]
[BlackType "program"]

1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.d4 Nxd4 6.Nxd4 exd4 7.O-O b5 8.Bb3 d6
9.Bd5 Rb8 10.Qxd4 Be7 11.Nc3 {+0.17/13 8} c5 {-0.22/11 9} 12.Qd3 {+0.22/14 8} O-O
{-0.22/10 9} 13.a3 {+0.26/14 8} Nxd5 {-0.06/12 8} 14.Nxd5 {+0.41/14 10} Bf6 {-0.06/9 8}
15.c3 {+0.33/13 10} Re8 {-0.06/9 9} 16.Bf4 {+0.38/14 10} Be5 {+0.02/9 9} 17.Bxe5
{+0.36/14 10} dxe5 {+0.04/9 8} 18.Qg3 {+0.41/14 10} f5 {+0.10/12 9} 19.f3 {+0.57/13 9} c4
{-0.02/11 9} 20.Rad1 {+0.34/13 9} Rb7 {+0.04/11 9} 21.exf5 {+0.42/14 12} Bxf5 {+0.08/9 7}
22.Rfe1 {+0.41/15 11} Qd6 {+0.08/7 7} 23.f4 {-0.23/14 11} e4 {+0.26/7 9} 24.Ne3
{+0.42/14 11} Qf6 {+0.10/7 9} 25.Rd5 {+0.46/14 10} Be6 {+0.10/7 8} 26.Rd6 {+0.86/13 11}
Rd7 {-0.42/7 7} 27.Rxa6 {+0.84/13 11} Qd8 {-0.84/7 8} 28.f5 {+1.41/14 12} Bf7
{-0.90/9 6} 29.f6 {+1.71/14 13} Bg6 {-0.92/9 6} 30.Qg5 {+1.63/14 12} Qb8 {-1.02/8 6}
31.fxg7 {+1.65/13 13} Rxg7 {-1.02/7 7} 32.Rf1 {+1.81/13 13} Kh8 {-1.34/8 7} 33.Ra5
{+1.81/11 14} Rb7 {-1.18/7 6} 34.Qf6+ {+2.39/5 14} Rg7 {-1.68/12 5} 35.a4 {+2.01/1 14}
bxa4 {-1.54/7 5} 36.Rxa4 {+1.55/12 15} Reg8 {-1.82/9 7} 37.Nxc4 {+1.64/12 15} Bf7
{-1.76/9 5} 38.Ne5 {+2.39/12 16} Bh5 {-1.56/8 5} {0-1 les Blancs perdent au temps} 0-1

[Event "Tournoi d'entrée"]
[Site "Le Fou numerique"]
[Date "2009.11.19"]
[Round "1"]
[White "Brutus 8.03"]
[Black "Aice 0.99.2"]
[Result "0-1"]
[ECO "A08"]
[Opening "Reti"]
[Time "20:02:52"]
[Variation "KIA, 2...c5 3.Bg2 Nc6 4.d4"]
[TimeControl "300+1"]
[Termination "time forfeit"]
[PlyCount "76"]
[WhiteType "program"]
[BlackType "program"]

1.Nf3 c5 2.g3 d5 3.d4 Nc6 4.Bg2 cxd4 5.Nxd4 e5 6.Nxc6 bxc6 7.c4 Be6 8.Nc3 e4
9.Qa4 Bd7 10.O-O Nf6 {-0.90/9 8} 11.Bg5 {+0.77/14 8} Be7 {-0.93/9 5} 12.Rad1
{+0.94/13 8} O-O {-1.03/9 6} 13.Qc2 {+1.09/14 8} Qc8 {-1.28/9 4} 14.cxd5 {+1.09/15 10}
cxd5 {-1.30/9 1} 15.Qb1 {+1.24/15 10} Bh3 {-1.74/10 8} 16.Bxf6 {+1.20/15 10} Bxf6
{-1.71/10 1} 17.Bxh3 {+1.29/15 10} Qxh3 {-1.53/12 6} 18.Nxd5 {+1.09/15 10} Qf5 {-1.69/11
53} 19.f3 {+1.46/13 9} Bg5 {-1.74/10 6} 20.Qxe4 {+1.91/14 9} Qxe4 {-1.56/12 7}
21.fxe4 {+1.83/16 9} Rab8 {-1.54/11 5} 22.b3 {+1.84/15 11} Rbc8 {-1.62/11 7} 23.h4
{+2.10/16 11} Bd8 {-1.74/13 5} 24.Rdc1 {+2.13/17 11} Bb6+ {-1.75/13 17} 25.e3 {+2.12/17
11} Bc5 {-1.41/11 4} 26.b4 {+2.13/16 11} Bd6 {-1.37/11 3} 27.Rxc8 {+1.91/15 11}
Rxc8 {-1.34/11} 28.e5 {+1.91/15 11} Bf8 {-1.20/13 11} 29.a3 {+1.63/15 13} Rc6
{-1.19/12 8} 30.Kf2 {+1.51/17 12} Re6 {-1.22/13 11} 31.Rd1 {+1.55/16 12} Rxe5 {-1.25/12
6} 32.Nf4 {+1.53/15 13} f6 {-1.06/11 6} 33.Rd7 {+1.43/13 14} Re7 {-1.00/12 5}
34.Rd8 {+1.43/6 14} Kf7 {-1.00/12 12} 35.g4 {+1.68/1 14} a5 {-1.07/11 7} 36.bxa5
{+1.50/14 15} Ra7 {-0.93/12 8} 37.Rd5 {+1.50/15 15} Bxa3 {-0.96/12 5} 38.Nd3 {+1.44/14
16} Be7 {-1.02/11 8} {0-1 les Blancs perdent au temps} 0-1

[Event "Tournoi d'entrée"]
[Site "Le Fou numerique"]
[Date "2009.11.19"]
[Round "1"]
[White "Brutus 8.03"]
[Black "Alaric 707"]
[Result "0-1"]
[ECO "E61"]
[Opening "King's Indian"]
[Time "20:13:14"]
[Variation "Fianchetto, Flohr Variation"]
[TimeControl "300+1"]
[Termination "time forfeit"]
[PlyCount "76"]
[WhiteType "program"]
[BlackType "program"]

1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.g3 d6 5.Bg2 O-O 6.e3 c5 7.Nge2 Nc6 8.O-O Bd7 9.b3
a6 10.Bb2 Rb8 11.dxc5 {+0.20/12 8} dxc5 {-0.14/15 11} 12.Nd5 {+0.26/12 8} Nxd5
{+0.62/18 12} 13.Bxg7 {-0.44/16 8} Nxe3 {+0.87/17 11} 14.fxe3 {-0.54/15 10} Kxg7
{+0.87/15 10} 15.Qc2 {-0.52/13 10} Qc7 {+0.93/14 12} 16.Nf4 {-0.61/14 10} Bf5 {+0.94/14
9} 17.Qf2 {-0.45/14 10} Ne5 {+0.97/13 9} 18.e4 {-0.29/14 10} Bg4 {+1.05/15 9}
19.h3 {-0.32/14 9} Be6 {+1.03/15 8} 20.Rad1 {-0.38/13 9} b5 {+1.14/15 8} 21.Nd5
{-0.50/13 12} Qa7 {+1.11/15 8} 22.cxb5 {-0.50/12 11} axb5 {+1.18/14 7} 23.Nf4 {-0.50/13
11} Qc7 {+1.40/13 7} 24.Nd5 {-0.50/13 11} Bxd5 {+1.45/16 7} 25.Rxd5 {-0.78/14 11}
c4 {+1.55/16 9} 26.Rc5 {-0.89/13 11} Qd6 {+1.63/16 7} 27.Rd5 {-0.91/13 11} Qe6
{+1.57/15 6} 28.Qc5 {-1.01/12 12} f6 {+1.73/13 6} 29.Kh1 {-1.08/13 13} Rfc8 {+2.08/15 6}
30.Qb4 {-0.94/14 12} c3 {+2.08/15 7} 31.Rc1 {-1.65/15 13} Nc6 {+2.38/16 7} 32.Qxc3
{-2.13/14 13} Nb4 {+2.45/16 5} 33.Qxb4 {-1.83/10 13} Rxc1+ {+2.64/15 5} 34.Kh2 {-1.83/7
14} Rc2 {+2.78/15 5} 35.Rd2 {-1.66/1 14} Rbc8 {+3.22/15 5} 36.Qd4 {-2.72/14 15} Qd6
{+3.07/16 10} 37.Qxd6 {-3.26/15 15} exd6 {+3.50/20 5} 38.Rxc2 {-3.30/16 16} Rxc2
{+3.49/19 4} {0-1 les Blancs perdent au temps} 0-1

Regards,
Patrick

[Stephan Vermeire]

Thanx for noticing the problem Patrick. The previous losts on time were during the very first move. These are not!
Could you tell me what GUI you are using? How many CPU's? I suppose you are testing under UCI? Any additional info is welcome.

Thanks already
Stephan

[Patrick Buchmann]

Thanx for noticing the problem Patrick. The previous losts on time were during the very first move. These are not!
Could you tell me what GUI you are using? How many CPU's? I suppose you are testing under UCI? Any additional info is welcome.

Thanks already
Stephan

Hi Stephan,

As usual UCI
5'+1", Quad Q9400 2.66Ghz, 2Go RAM, Win Vista SP2, Arena 1.1 GUI
Hash 128Mo, Biblio HS-Book.abk, Tablebases 5-pièces

The version 8.01 and 8.02 don't loose on time after the book moves, but as 8.03 spend too much time (5-8 seconds) for begin to comput. So they lost many seconds on each move and the 5-8 seconds pause in the last move is the cause for the definitif lost.

Regards,
Patrick

[Stephan Vermeire]

Hi Stephan,

As usual UCI
5'+1", Quad Q9400 2.66Ghz, 2Go RAM, Win Vista SP2, Arena 1.1 GUI
Hash 128Mo, Biblio HS-Book.abk, Tablebases 5-pièces

The version 8.01 and 8.02 don't loose on time after the book moves, but as 8.03 spend too much time (5-8 seconds) for begin to comput. So they lost many seconds on each move and the 5-8 seconds pause in the last move is the cause for the definitif lost.

Regards,
Patrick

Hi Patrick,

Thanx for the info. I will try to reproduce the problem on my own computer. Personally I am usually not running under Arena with egtb= 5 pieces, that might have something to do with it. If I cannot reproduce the problem, I will contact you again OK?

Best wishes,
Stephan

[Graham Banks]

Brutus 8.03b lost its first game on time here under ChessGUI too.
Here is the debug file from ChessGUI.
http://www.kirill-kryukov.com/chess/dis ... p?id=16879

Cheers,
Graham.

[Stephan Vermeire]

Hi Graham,

Great that you have a debug-log! I found the problem within seconds: Under UCI Brutus is setting up the position and is printing the board after eacht move. this is a time-consuming business! Especially a bit further on in the game, Brutus needed many ms and lost the game as a result. Easy to fix! Thanx Graham and Patrick!

I also found another bug already that occurs at the very first move in multi-CPU. It appeared that the actual thinking started before the processes were fully initiated. I hope that it has been fixed with a small adjustment. I am going to test it here for a few hundered games and an upgrade will be posted very soon if all goes well.

Best wishes,
Stephan

[Graham Banks]

Hi Graham,

Great that you have a debug-log! I found the problem within seconds: Under UCI Brutus is setting up the position and is printing the board after eacht move. this is a time-consuming business! Especially a bit further on in the game, Brutus needed many ms and lost the game as a result. Easy to fix! Thanx Graham and Patrick!

I also found another bug already that occurs at the very first move in multi-CPU. It appeared that the actual thinking started before the processes were fully initiated. I hope that it has been fixed with a small adjustment. I am going to test it here for a few hundered games and an upgrade will be posted very soon if all goes well.

Best wishes,
Stephan

Sounds good.

[Patrick Buchmann]

If I cannot reproduce the problem, I will contact you again OK?

Best wishes,
Stephan

Hi Stephan,

It seems the problem occurs on each move.

Befor the engine calculate, it sends ucinewgame, itsready and readyok. So for each move it restarts as was a new game and lost 5 to 8 seconds pro move.

Regards,
Patrick