OliThink 5.2.2 released with 48MB Hashsize

[Graham Banks]

Very first game:


[Event "Gauntlet"]
[Site "ChessGUI"]
[Date "2009.12.08"]
[Time "5:06:57 p.m."]
[Round "1"]
[White "Gibbon 2.52a"]
[Black "OliThink 5.2.2"]
[Result "1-0"]
[TimeControl "40/3720:40/3720:40/3720"]
[PlyCount "52"]
[Number "1"]
[BlackType "program"]
[WhiteType "program"]
[Variant "normal"]

{ Unknown Hardware }
1.e4 {[%eval 0,1] [%emt 00:00:00]} e5 {[%eval 0,1] [%emt 00:00:00]} 2.Nf3 {[%eval 0,1] [%emt 00:00:00]} Nc6 {[%eval 0,1] [%emt 00:00:00]}
3.Bb5 {[%eval 0,1] [%emt 00:00:00]} a6 {[%eval 0,1] [%emt 00:00:00]} 4.Ba4 {[%eval 0,1] [%emt 00:00:00]} Nf6 {[%eval 0,1] [%emt 00:00:00]}
5.O-O {[%eval 0,1] [%emt 00:00:00]} Be7 {[%eval 0,1] [%emt 00:00:00]} 6.Re1 {[%eval 0,1] [%emt 00:00:00]} b5 {[%eval 0,1] [%emt 00:00:00]}
7.Bb3 {[%eval 0,1] [%emt 00:00:00]} d6 {[%eval 0,1] [%emt 00:00:00]} 8.c3 {[%eval 0,1] [%emt 00:00:00]} O-O {[%eval 0,1] [%emt 00:00:00]}
9.h3 {[%eval 0,1] [%emt 00:00:00]} Bb7 {[%eval 0,1] [%emt 00:00:00]} 10.d4 {[%eval 0,1] [%emt 00:00:00]} Na5 {[%eval 0,1] [%emt 00:00:00]}
11.Bc2 {[%eval 0,1] [%emt 00:00:00]} Nc4 {[%eval 0,1] [%emt 00:00:00]}
12.b3 {[%eval 0,1] [%emt 00:00:00]} Nb6 {[%eval 0,1] [%emt 00:00:00]} 13.d5 {[%eval 4,12] [%emt 00:04:18]} Nfd7 {[%eval -17,12] [%emt 00:02:06]}
14.Qe2 {[%eval 20,12] [%emt 00:02:01]} c6 {[%eval -6,13] [%emt 00:01:21]}
15.dxc6 {[%eval -2,14] [%emt 00:05:28]} Bxc6 {[%eval 0,13] [%emt 00:01:24]}
16.Be3 {[%eval -2,13] [%emt 00:05:25]} Qc7 {[%eval -9,12] [%emt 00:02:12]}
17.g3 {[%eval 2,10] [%emt 00:01:04]} Rfc8 {[%eval 14,11] [%emt 00:01:28]}
18.Kh2 {[%eval 12,11] [%emt 00:01:53]} d5 {[%eval 33,12] [%emt 00:02:13]}
19.Bxb6 {[%eval 6,11] [%emt 00:01:02]} Qxb6 {[%eval 53,13] [%emt 00:02:13]}
20.exd5 {[%eval 19,12] [%emt 00:04:31]} Bxd5 {[%eval 44,12] [%emt 00:02:13]}
21.a4 {[%eval -1,11] [%emt 00:01:07]} bxa4 {[%eval 86,11] [%emt 00:01:19]}
22.Rxa4 {[%eval -1,10] [%emt 00:00:50]} Qd4 {[%eval -969,11] [%emt 00:09:05]}
23.Nxd4 {[%eval 1044,11] [%emt 00:00:45]} Bc5 {[%eval -981,11] [%emt 00:01:08]}
24.Nf5 {[%eval 1113,12] [%emt 00:01:50]} g6 {[%eval -1032,12] [%emt 00:01:57]}
25.Rxa6 {[%eval 1165,12] [%emt 00:02:58]} Rxa6 {[%eval -1052,13] [%emt 00:01:57]}
26.Qxa6 {[%eval 1165,12] [%emt 00:00:25]} Ra8 {[%eval -1054,11] [%emt 00:01:11]}
1-0
{ Game Nr. 1 : Gibbon 2.52a wins against OliThink 5.2.2 by GUI adjudication
GameDuration = 01:05:45, Finalposition r5k1/3n1p1p/Q5p1/2bbpN2/8/1PP3PP/2B2P1K/1N2R3 w - - 1 27 }

Debug file available from the CCRL public forum.
http://www.kirill-kryukov.com/chess/dis ... p?id=17110

[Guenther]

Very first game:


[Event "Gauntlet"]
[Site "ChessGUI"]
[Date "2009.12.08"]
[Time "5:06:57 p.m."]
[Round "1"]
[White "Gibbon 2.52a"]
[Black "OliThink 5.2.2"]
[Result "1-0"]
[TimeControl "40/3720:40/3720:40/3720"]
[PlyCount "52"]
[Number "1"]
[BlackType "program"]
[WhiteType "program"]
[Variant "normal"]

{ Unknown Hardware }
1.e4 {[%eval 0,1] [%emt 00:00:00]} e5 {[%eval 0,1] [%emt 00:00:00]} 2.Nf3 {[%eval 0,1] [%emt 00:00:00]} Nc6 {[%eval 0,1] [%emt 00:00:00]}
3.Bb5 {[%eval 0,1] [%emt 00:00:00]} a6 {[%eval 0,1] [%emt 00:00:00]} 4.Ba4 {[%eval 0,1] [%emt 00:00:00]} Nf6 {[%eval 0,1] [%emt 00:00:00]}
5.O-O {[%eval 0,1] [%emt 00:00:00]} Be7 {[%eval 0,1] [%emt 00:00:00]} 6.Re1 {[%eval 0,1] [%emt 00:00:00]} b5 {[%eval 0,1] [%emt 00:00:00]}
7.Bb3 {[%eval 0,1] [%emt 00:00:00]} d6 {[%eval 0,1] [%emt 00:00:00]} 8.c3 {[%eval 0,1] [%emt 00:00:00]} O-O {[%eval 0,1] [%emt 00:00:00]}
9.h3 {[%eval 0,1] [%emt 00:00:00]} Bb7 {[%eval 0,1] [%emt 00:00:00]} 10.d4 {[%eval 0,1] [%emt 00:00:00]} Na5 {[%eval 0,1] [%emt 00:00:00]}
11.Bc2 {[%eval 0,1] [%emt 00:00:00]} Nc4 {[%eval 0,1] [%emt 00:00:00]}
12.b3 {[%eval 0,1] [%emt 00:00:00]} Nb6 {[%eval 0,1] [%emt 00:00:00]} 13.d5 {[%eval 4,12] [%emt 00:04:18]} Nfd7 {[%eval -17,12] [%emt 00:02:06]}
14.Qe2 {[%eval 20,12] [%emt 00:02:01]} c6 {[%eval -6,13] [%emt 00:01:21]}
15.dxc6 {[%eval -2,14] [%emt 00:05:28]} Bxc6 {[%eval 0,13] [%emt 00:01:24]}
16.Be3 {[%eval -2,13] [%emt 00:05:25]} Qc7 {[%eval -9,12] [%emt 00:02:12]}
17.g3 {[%eval 2,10] [%emt 00:01:04]} Rfc8 {[%eval 14,11] [%emt 00:01:28]}
18.Kh2 {[%eval 12,11] [%emt 00:01:53]} d5 {[%eval 33,12] [%emt 00:02:13]}
19.Bxb6 {[%eval 6,11] [%emt 00:01:02]} Qxb6 {[%eval 53,13] [%emt 00:02:13]}
20.exd5 {[%eval 19,12] [%emt 00:04:31]} Bxd5 {[%eval 44,12] [%emt 00:02:13]}
21.a4 {[%eval -1,11] [%emt 00:01:07]} bxa4 {[%eval 86,11] [%emt 00:01:19]}
22.Rxa4 {[%eval -1,10] [%emt 00:00:50]} Qd4 {[%eval -969,11] [%emt 00:09:05]}
23.Nxd4 {[%eval 1044,11] [%emt 00:00:45]} Bc5 {[%eval -981,11] [%emt 00:01:08]}
24.Nf5 {[%eval 1113,12] [%emt 00:01:50]} g6 {[%eval -1032,12] [%emt 00:01:57]}
25.Rxa6 {[%eval 1165,12] [%emt 00:02:58]} Rxa6 {[%eval -1052,13] [%emt 00:01:57]}
26.Qxa6 {[%eval 1165,12] [%emt 00:00:25]} Ra8 {[%eval -1054,11] [%emt 00:01:11]}
1-0
{ Game Nr. 1 : Gibbon 2.52a wins against OliThink 5.2.2 by GUI adjudication
GameDuration = 01:05:45, Finalposition r5k1/3n1p1p/Q5p1/2bbpN2/8/1PP3PP/2B2P1K/1N2R3 w - - 1 27 }

Debug file available from the CCRL public forum.
http://www.kirill-kryukov.com/chess/dis ... p?id=17110

Graham one question:
Was this the original 5.22 version or the JA version?

Guenther

[Graham Banks]

Graham one question:
Was this the original 5.22 version or the JA version?

Guenther

The JA version playing as a Winboard engine under ChessGUI.

Cheers,
Graham.

[Guenther]

Graham one question:
Was this the original 5.22 version or the JA version?

Guenther

The JA version playing as a Winboard engine under ChessGUI.

Cheers,
Graham.

Well, then you should use the original version of Olithink, because it
most probably doesn't contain this bug. It seems this is another
manifestation of the bug introduced by compiling with MSVC.
I never have seen a giveaway bug or illegal Q move bug in the original
5.xx versions tested here.
Note that I am still testing a 5.21iJA(Intel compile) for Jim for this reason.
A bit more info in this link:
http://www.open-aurec.com/wbforum/viewt ... ug#p191903

Guenther

[Graham Banks]

you should use the original version of Olithink, because it most probably doesn't contain this bug.

Guenther

Is there a 32-bit compile available for windows from Oliver?

Cheers,
Graham.

[swami]

you should use the original version of Olithink, because it most probably doesn't contain this bug.

Guenther

Is there a 32-bit compile available for windows from Oliver?

Cheers,
Graham.

http://home.arcor.de/dreamlike/chess/Ol ... ingw32.zip

[Graham Banks]

you should use the original version of Olithink, because it most probably doesn't contain this bug.

Guenther

Is there a 32-bit compile available for windows from Oliver?

Cheers,
Graham.

http://home.arcor.de/dreamlike/chess/Ol ... ingw32.zip

Thanks Swami. Not sure how I missed it.

Cheers,
Graham.

[OliverBr]

Graham one question:
Was this the original 5.22 version or the JA version?

Guenther

The JA version playing as a Winboard engine under ChessGUI.

Cheers,
Graham.

Ups! You shoudln't use that one.

Look at my post from September:

http://www.talkchess.com/forum/viewtopi ... t&start=10

Whatever version/compile of OiThink 5.2.2 must yield the following log for the first 10 depths when answering to 1. e4 without book:

Code: Select all

# ./olithink64 -sd 10
e4
 1     1      0        30  e7e6
 2   -19      0       242  d7d5 d2d3
 3    -1      0       906  d7d5 b1c3 g8f6
 4   -16      0      2756  d7d5 b1c3 d5e4 c3e4
 5     9      2     12000  d7d5 e4e5 b8c6 d2d4 c8f5
 6   -25     11     59422  d7d5 e4e5 e7e6 d2d4 f8b4 c2c3 b4e7
 7    -6     69    355080  e7e6 g1f3 c7c6 b1c3 f8d6 e4e5 d6c5
 8   -26    179   1032725  e7e5 f1c4 g8f6 b1c3 f8c5 g1f3 b8c6 d2d3
 9   -11    325   3278020  d7d5 e4d5 d8d5 b1c3 d5e6 d1e2 g8f6 d2d3 e6e2 f1e2 b8c6
10   -21    667   8698717  d7d5 e4e5 d5d4 f1c4 b8c6 g1f3 e7e6 a2a4 f8b4 c2c3
1. ... d7d5

kibitz W: -21 Nodes: 6220053 QNodes: 2478664 Evals: 4025445 cs: 667 knps: 1303

You can do this in windows. too, using cmd.
There mustn't be any difference, not one single digit. It doesn't matter if 64bit or 32bit compile, Linux or Windows, Java or native.... It's a chess engine and not a random generator

PS: Just to see that there is something wrong with the Jim Ablett compile:

Code: Select all

C:\Programme\WinBoard-4.2.7>olithink222-ja -sd 10
e4
 1     1      0        30  e7e6
 2   -19      0       251  d7d5 d2d3
 3    -1      0       918  d7d5 b1c3 g8f6
 4   -16      0      3010  d7d5 b1c3 d5e4 c3e4
 5     9      1     13681  d7d5 e4e5 b8c6 d2d4 c8f5
 6   -25      7     65623  d7d5 e4e5 e7e6 d2d4 f8b4 c2c3 b4e7
 7    -7     35    348084  e7e6 g1f3 b7b6 d2d4 g8f6 f1d3 f8d6
 8   -26    123   1217736  e7e5 b1c3 g8f6 g1f3 b8c6 f1c4 f8c5 d2d3
 9   -11    514   5265027  d7d5 e4d5 d8d5 b1c3 d5e6 d1e2 g8f6 d2d3 e6e2 f1e2 b8c6
10   -21   1192  12347116  d7d5 e4e5 d5d4 f1c4 b8c6 g1f3 e7e6 a2a4 f8b4 c2c3
1. ... d7d5

kibitz W: -21 Nodes: 8980170 QNodes: 3366946 Evals: 5218731 cs: 1192 knps: 1035

It's very different and thus wrong. Looks like the move generator is different, there are more nodes and those illegal moves.

[Jim Ablett]

I've removed the buggy Msvc compiles from my homepage.
I'll re-compile with the Intel compiler if it proves not to introduce this bug.

Jim.

[OliverBr]

I've removed the buggy Msvc compiles from my homepage.
I'll re-compile with the Intel compiler if it proves not to introduce this bug.

Jim.

Thank you very much, Jim, for all your great work. Anyway, it should work with Msvc, too, as I used to use Msvc. It would be interesting to find out, what happens now.