Glaurung 2.0.1 Windows build available

[Werner]

Glaurung 2.0.1 (chessbase fix) win32 Intel compiler 10 (p.g.o)

http://www.zshare.net/download/5129511eccb5d2/
Mirrors:
http://mihd.net/eltghn
http://w14.easy-share.com/10959611.html

Jim

Hi Jim,
thanks a lot - its working ok inside Fritz 11.
But I saw - you first version with 847 kB is a bit faster!

regards

[Jim Ablett]

Hi Werner

thanks a lot - its working ok inside Fritz 11.
But I saw - you first version with 847 kB is a bit faster!

Strange. Speed should be identical - I used exactly the same compiler settings/optimizations.

Jim.

[Werner]

[quote="WernerHi Jim,
thanks a lot - its working ok inside Fritz 11.
But I saw - you first version with 847 kB is a bit faster!

regards[/quote]

...as we know the reason of the crashes - I edited your first version with a Hex-editor and instead of the text
Glaurung is free software, and you are welcome to redistribute it under certain conditions; see the file Copying.txt for details.
I used 00 for each character. Now I can load this version too into Fritz 11

I have only small differences in starting position: 1241 kns ->1276 and this results in 1-2 seconds better to reach T18 and 3 seconds for T19.
Now conformed in Chessbase and Arena GUI after each 2 rounds....

[Jim Ablett]

edited your first version with a Hex-editor

Could you put a download link to it here, I'd like to test it.

Jim.

[mariaclara]

..

Thanks Jim, Tord...

[Werner]

edited your first version with a Hex-editor

Could you put a download link to it here, I'd like to test it.

Jim.

Hi Jim,
I went to your homepage and sent you the file as I have no server here.
The file is glaurung.zip

[Jim Ablett]

I did remove a compile switch while bug-hunting the opening book problem and forgot to put it back.
The /Qparallel compiler switch auto-parallelizes the loops in the code.

New compile to test (hopefully back up to speed now)

http://www.zshare.net/download/513338758ace3e/

Jim.

[David Dahlem]

My new logo for Glaurung 2.0.1 can be found here ...

http://www.geocities.com/davedahlem/my_logos.html

Regards
Dave

[Werner]

I did remove a compile switch while bug-hunting the opening book problem and forgot to put it back.
The /Qparallel compiler switch auto-parallelizes the loops in the code.

New compile to test (hopefully back up to speed now)

http://www.zshare.net/download/513338758ace3e/

Jim.

Hi Jim,
the file is much larger now and you are nearly back to old speed:

Analysis by Glaurung 2.0.1 w32 1CPU:

1.e4 e5 2.Sf3 Sc6 3.Lb5 Sf6 4.0-0 Lc5 5.Sc3 0-0 6.Sxe5 Sxe5 7.d4 Ld6 8.dxe5 Lxe5 9.Lg5 d6
+/= (0.29) Tiefe: 18 00:00:39 49617kN
1.e4 e5 2.Sf3 Sc6 3.Lb5 Sf6 4.0-0 Lc5 5.c3 0-0 6.d4 Le7 7.Te1 exd4 8.cxd4 d5 9.Lxc6 bxc6 10.Sc3 dxe4 11.Sxe4 Sd5
+/= (0.31) Tiefe: 19 00:01:12 90450kN


and old version:

Neue Partie - Glaurung 2.0.1we 32bit 1CPU, 11'/40+11'/40+11'/40
rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1

Analysis by Glaurung 2.0.1we 32bit 1CPU:

1.e4 e5 2.Sf3 Sc6 3.Lb5 Sf6 4.0-0 Lc5 5.Sc3 0-0 6.Sxe5 Sxe5 7.d4 Ld6 8.dxe5 Lxe5 9.Lg5 d6
+/= (0.29) Tiefe: 18 00:00:39 49617kN
1.e4 e5 2.Sf3 Sc6 3.Lb5 Sf6 4.0-0 Lc5 5.c3 0-0 6.d4 Le7 7.Te1 exd4 8.cxd4 d5 9.Lxc6 bxc6 10.Sc3 dxe4 11.Sxe4 Sd5
+/= (0.31) Tiefe: 19 00:01:10 90450kN

2 seconds....

[Eelco de Groot]

To be honest Tord, I do miss the adjustable Aggressiveness and Cowardice etc. from Glaurung 1 a bit I hope you don't mind me saying so... I hope they will return

I don't mind you saying so, and I also hope they will return some day. The reason they are not included in Glaurung 2.0 is that I need to write an evaluation function before I can offer any interesting user configurable evaluation parameters. The evaluation function in Glaurung 2.0 is still very simple and primitive. In particular, it knows almost nothing about king safety, which is what the Aggressiveness and Cowardice parameters in Glaurung 1 are about.

Tord

Thanks Tord,

Is it not strange that King Safety is such a double edged sword and it is very hard to increase the elo with making a better King Safety function?

I did notice the new Selectivity parameter you made in Glaurung, very interesting, and I was experimenting a bit with it. Tord I think it is possible the default options as they now stand may not be the best, maybe it is because I am testing the single core version or maybe it is because I am using Harry Schnapp's Rybka book for Glaurung, -it is very good but maybe Glaurung in the default setings is not making the most of it-, or maybe a combination of factors, but the results I get against the brand new Belka 1.8.11 are not so good at all, in the default settings.

I still have not played many games but the results with some variation of the "Mjolnir"' settings seemed to do better, against Belka at least.

Any chance that you would consider changing the defaults Tord? Normally I would not think there would be big difference with other settings but in this case I do think it would have a chance of getting better results.

What I have been testing is, I posted it on the Rybka forum as well:

Selectivity: 5 (1 step down from default)
Single Reply Extension (PV): 2 (1 up from default 1, 1 is a half ply I believe)
Single Reply Extension (non PV): 2 (1 up from default 1)

Results with the Arasan 7 test of feb 27 2007 were about equal:

Code: Select all

Glaurung 2.0.1 default  128 Mb HT 20 sec per pos

Right until now: 164 of 227  ;  32:01m


     1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------
  0| 0   4   0   3   -   0   0   4   5   2   0   1   0   1   0   -   0   0   0   1
 20| 4   4   7   1   -  13   0   -   2   1   1   0   0   6   0   1   1  18   1   -
 40| 0   1  11   8   0   -   1  12   2   0   0   0   -   0  10   0   0   -  12   0
 60| 0   -   1   -   -   0   -   -   1   5   1  17   -   0   6   0   2   9   -   7
 80| -   6   8   -   0   0   0   2   0   1   5   0   0   0   1   2  14   3   -   0
100| 0   0   -  15   7   6   0   0  18   1   -   7   3   -   3  17   -   8  10   0
120|13   -   1   0   1   -   2   5   4   -   -   4   0   1   -   1   -   0   -   2
140| 2  14   4   -   0   4   -   -  14   -   6   0   0   0   -  12   1  19  13   -
160| 6  13   -   -   -   4   5   0   -   0   0   2   -   -   -   -   -   -   6   -
180| 0   0   5   3  11   0   -   0  12   -   0   2   -   5   4   -   -   2   -   4
200| -   8   0   0   0   -   -  17   -  13   0   6   3   2   -   5   -   0   1   -
220| -   4   -   0   -   0   -

   1 sec ->  62/227
   2 sec ->  85/227
   3 sec ->  98/227
   4 sec -> 104/227
   5 sec -> 116/227
   6 sec -> 124/227
   7 sec -> 132/227
   8 sec -> 136/227
   9 sec -> 140/227
  10 sec -> 141/227
  11 sec -> 143/227
  12 sec -> 145/227
  13 sec -> 149/227
  14 sec -> 154/227
  15 sec -> 157/227
  20 sec -> 164/227
  n/s: 773.716

TotTime: 76:11m    SolTime: 32:01m

Code: Select all

Glaurung 2.0.1 Mjolnir 001b
Right until now: 165 of 227  ;  31:48m


    1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
----------------------------------------------------------------------------------  0|0   6   0   4   -   0   0   4   -   4   0   5   0   0   -   3   1   0   3   1
 20|4   -   7   3   -   8   1   -   8   4   1   0   0   5   0   2   4   -   0   4
 40|1   3  10   0   0   6   1  16   9   1   0   0  16   0   1   2   0   -   7   0
 60|2   -   1   -   -   0   -   -   1   6   1  12   -   0   2   0   2   -   -  11
 80|-   7   9   -   1   0   0   1   0   1   1   0   0   0   3   1   -   9   8   0
100|1   0   -  13  14   9   0   0   -   1   -   6   6   -   4   9   0   2  10   1
120|0   -   4   1   -   8   0   5   5   -   -   7   0   1  12   1   -   0   -   2
140|-   -   5   -   0   6  15   -  19   -   0   0   1   0   -   7   2  17   -   -
160|6   -   -   -   -   5   2   0   -   0   3   0   -   -  18   -   -   -   3   -
180|0   0   2   8   9   8   -   0  13   -   0   2   3   6   1   -   -   2   -   2
200|-   3   0   0   0   0   -   9  16   -   0   3   2   2   -   4   -   0   3   -
220|-   7   -   0   -   5  17

   1 sec ->  55/227
   2 sec ->  79/227
   3 sec ->  94/227
   4 sec -> 105/227
   5 sec -> 115/227
   6 sec -> 122/227
   7 sec -> 130/227
   8 sec -> 136/227
   9 sec -> 142/227
  10 sec -> 149/227
  11 sec -> 151/227
  12 sec -> 152/227
  13 sec -> 154/227
  14 sec -> 156/227
  15 sec -> 157/227
  20 sec -> 165/227
  n/s: 776.666 

TotTime: 76:08m    SolTime: 31:48m

As you can see not much of a change overall but in some positions 'not solved' is changed to 'solved very quickly' or vice versa so the move chosen is sometimes different it seems.

I also tested the Mjolnir settings with threat depth 4 (from 3) and at first the Arasan test seemed to go better but at the end, with some more difficult problems the end result was about the same, 164 positions, same as the default.

This is the result for the Lasker - Sterk position at longer analysis:

Glaurung 2.0.1 default on Aljechin - Sterk

[d]r4rk1/p4ppp/qp2p3/b5B1/n1R5/5N2/PP2QPPP/1R4K1 w - -

Engine: Glaurung 2.0.1 default (256 MB)
by Tord Romstad

2.00 0:00 -1.54 1.Rd1 Nxb2 (159) 0

2.00 0:00 -0.56 1.Ne5 Nc5 (219) 0

3.00 0:00 -0.52 1.Ne5 Rac8 2.Rbc1 (914) 2

3.00 0:00 -0.13 1.Qc2 b5 2.Rf4 (1.251) 3

4.00 0:00 -0.80 1.Qc2 b5 2.Rh4 h6 (2.058) 6

4.00 0:00 -0.58 1.Ne5 Rac8 2.Rbc1 Nc5 (2.660) 8

4.00 0:00 -0.33 1.Be7 Rfc8 2.Rxc8+ Qxc8 3.Qe4 Nc5 (3.220) 9

5.00 0:00 -0.37 1.Be7 Rfc8 2.Rxc8+ Qxc8 3.Ne5 Qb7
4.Bd6 (4.980) 15

6.00 0:00 -0.52 1.Be7 Rfc8 2.Rxc8+ Qxc8 3.Ne5 Qb7
4.Bd6 Nc5 (8.012) 23

6.00 0:00 -0.01 1.Qc2 b5 2.Rh4 h6 3.Be7 Rac8 (10.507) 30

7.00 0:00 -0.27 1.Qc2 b5 2.Rh4 h6 3.Be7 Rac8 4.Qd3 (14.698) 40

8.00 0:00 -0.37 1.Qc2 b5 2.Rc6 Qb7 3.Rc1 f6 4.Be3 e5
5.Qb3+ Rf7 (30.697) 81

8.00 0:00 +1.00 1.Qf1 Nc5 2.b4 Bxb4 3.Rbxb4 Qxa2
4.Ne5 Rac8 (68.625) 163

9.00 0:00 +1.11 1.Qf1 Nc5 2.b4 Bxb4 3.Rcxb4 Qxa2
4.Be7 Rfc8 5.Ne5 (98.184) 209

10.00 0:00 +1.39 1.Qf1 Nc5 2.b4 Bxb4 3.Rbxb4 Qxa2
4.Be7 Rfc8 5.Ne5 Rc7 (147.512) 277

11.00 0:00 +1.17 1.Qf1 Nc5 2.b4 Bxb4 3.Rbxb4 Qxa2
4.Be7 Rfc8 5.Rb5 Nb3 6.Rf4 (275.733) 401

12.01 0:01 +1.41 1.Qf1 Nc5 2.b4 Bxb4 3.Rcxb4 Qxa2
4.Be7 Rfc8 5.Ne5 Rc7 6.Bd6 Rcc8 (784.035) 583

13.01 0:02 +1.29 1.Qf1 Nc5 2.b4 Bxb4 3.Rcxb4 Qxa2
4.Be7 Rfc8 5.Ne5 f6 6.Ra1 Qc2 7.Rc4 (1.497.334) 665

14.01 0:04 +0.98 1.Qf1 Nc5 2.b4 Bxb4 3.Rcxb4 Qxa2
4.Be7 Rfc8 5.Rg4 f5 6.Rh4 Ne4 7.Ng5 Nxg5
8.Bxg5 (3.184.108) 735

14.02 0:06 +1.76 1.Nd4 Nc5 2.b4 Bxb4 3.Rbxb4 Rac8
4.Rb2 Qb7 5.Rbc2 f6 6.Be3 Rfd8 7.Nb5 Qe7 (5.126.656) 752

15.01 0:12 +1.84 1.Nd4 Nc5 2.b4 Bxb4 3.Rbxb4 Rac8
4.Rb2 Qb7 5.Rbc2 f6 6.Be3 Rfd8 7.Nb3 Nxb3
8.axb3 Rxc4 9.Rxc4 (9.537.883) 778

16.01 0:21 +2.00 1.Nd4 Nc5 2.b4 Bxb4 3.Rbxb4 Rac8
4.Rb2 Qb7 5.Rbc2 f6 6.Be3 Rfd8 7.Qg4 f5
8.Qh5 Rc7 (17.060.826) 790

17.01 0:42 +1.90 1.Nd4 Nc5 2.b4 Ne4 3.Bf4 Rad8 4.Qc2 b5
5.Rc6 Bb6 6.Qxe4 Qxa2 7.Ne2 Qd5
8.Nc3 Qxe4 9.Nxe4 Bd4 (33.962.692) 799

18.01 1:33 +1.82 1.Nd4 Nc5 2.b4 Ne4 3.Bf4 Rad8 4.Qc2 b5
5.Rc6 Bb6 6.Qxe4 Qxa2 7.Ne2 Rfe8
8.Bd6 Qd5 9.Qxd5 exd5 10.Nf4 (75.211.664) 801

18.04 5:05 +3.11 1.Bf6 Rfc8 2.Qe5 Rc5 3.Qg3 g6 4.Qh4 h5
5.Rxa4 Rac8 6.Rd4 Rc1+ 7.Rd1 Rxb1
8.Rxb1 Qd3 9.Ra1 Qc2 10.Qf4 a6 (242.059.463) 791

19.01 8:11 +4.13 1.Bf6 Rfc8 2.Qe5 Rc5 3.Rxc5 gxf6
4.Qg3+ Kh8 5.Ng5 fxg5 6.Qe5+ Kg8
7.Qxg5+ Kf8 8.Rc7 Qd3 9.Qe7+ Kg7
10.Qxf7+ Kh6 11.Qxe6+ Qg6 12.Qxg6+ Kxg6
13.b4 Bxb4 14.Rxb4 (391.175.805) 796

20.01 11:04 +4.01 1.Bf6 Rfc8 2.Qe5 Rxc4 3.Qg5 Rg4
4.Qxg4 g6 5.Qxa4 b5 6.Qh4 Qd6 7.Ne5 h5
8.Qg3 Rc8 9.Nxg6 Qxg3 10.Ne7+ Kf8
11.hxg3 (531.593.326) 799

21.01 19:48 +4.33 1.Bf6 Rfc8 2.Qe5 Rxc4 3.Qg5 Rg4
4.Qxg4 g6 5.Qxa4 b5 6.Qa3 b4 7.Qa4 Qd3
8.Rc1 Qd5 9.Ne5 Qd2 10.Rd1 Qf4
11.Qc6 (950.562.224) 799

best move: Bg5-f6 time: 24:41.985 min n/s: 802.119 nodes: 1.188.690.000

r4rk1/p4ppp/qp2p3/b5B1/n1R5/5N2/PP2QPPP/1R4K1 w - -

Engine: Glaurung 2.0.1 Mjolnir 001b (256 MB)
by Tord Romstad

2.00 0:00 -1.54 1.Rd1 Nxb2 (159) 0

2.00 0:00 -0.56 1.Ne5 Nc5 (219) 0

3.00 0:00 -0.52 1.Ne5 Rac8 2.Rbc1 (914) 3

3.00 0:00 -0.13 1.Qc2 b5 2.Rf4 (1.253) 4

4.00 0:00 -0.80 1.Qc2 b5 2.Rh4 h6 (2.060) 7

4.00 0:00 -0.58 1.Ne5 Rac8 2.Rbc1 Nc5 (2.662) 9

4.00 0:00 -0.33 1.Be7 Rfc8 2.Rxc8+ Qxc8 3.Qe4 Nc5 (3.222) 11

5.00 0:00 -0.37 1.Be7 Rfc8 2.Rxc8+ Qxc8 3.Ne5 Qb7
4.Bd6 (4.987) 16

6.00 0:00 -0.52 1.Be7 Rfc8 2.Rxc8+ Qxc8 3.Ne5 Qb7
4.Bd6 Nc5 (8.027) 27

6.00 0:00 -0.01 1.Qc2 b5 2.Rh4 h6 3.Be7 Rac8 (10.648) 35

6.00 0:00 +0.29 1.Nd4 Nc5 2.b4 h6 3.Be3 e5 4.bxc5 exd4
5.Bxd4 (14.119) 45

7.00 0:00 +0.72 1.Nd4 Nc5 2.b4 h6 3.Be3 e5 4.Rxc5 Qxe2
5.Nxe2 bxc5 6.bxa5 (21.030) 67

7.00 0:00 +1.01 1.Qf1 Nc5 2.b4 Bxb4 3.Rbxb4 Qxa2
4.Ne5 (30.498) 92

8.00 0:00 +1.00 1.Qf1 Nc5 2.b4 Bxb4 3.Rbxb4 Qxa2
4.Ne5 Rac8 (46.346) 128

9.00 0:00 +1.09 1.Qf1 Nc5 2.b4 Bxb4 3.Rbxb4 Qxa2
4.Ne5 Rac8 5.Be7 (83.186) 204

9.00 0:00 +1.35 1.Nd4 Nc5 2.b4 e5 3.Nf5 f6 4.bxc5 fxg5
5.Qxe5 (131.239) 279

10.00 0:00 +1.39 1.Nd4 Nc5 2.b4 e5 3.Nf5 f6 4.Bh6 gxh6
5.Qg4+ Kf7 6.Nxh6+ Ke7 7.bxc5 (184.164) 346

11.00 0:00 +0.96 1.Nd4 Nc5 2.b4 Ne4 3.Be7 Rfe8 4.bxa5 Nd2
5.Qxd2 Qxc4 6.Rc1 (401.421) 493

11.07 0:01 +1.17 1.Qf1 Nc5 2.b4 Bxb4 3.Rbxb4 Qxa2
4.Be7 Rfc8 5.Rb5 Nb3 6.Rf4 (752.147) 594

12.01 0:01 +1.39 1.Qf1 Nc5 2.b4 Bxb4 3.Rcxb4 Qxa2
4.Be7 Rfe8 5.Bxc5 bxc5 6.R1b2 Qd5
7.Rb7 (965.779) 637

13.01 0:02 +1.25 1.Qf1 Nc5 2.b4 Bxb4 3.Rcxb4 Qxa2
4.Rg4 f6 5.Bh6 Rf7 6.Ra1 Qc2 7.Nd4 (1.799.966) 706

14.01 0:06 +0.96 1.Qf1 Nc5 2.b4 Bxb4 3.Rcxb4 Qxa2
4.Rg4 f6 5.Bh6 Rf7 6.Ra1 Qb2 7.Rd1 g5 (4.666.229) 765

14.02 0:07 +1.52 1.Kf1 f6 2.Be3 Qb5 3.b4 Rac8 4.Rxc8 Qxe2+
5.Kxe2 Rxc8 6.Bd2 Bxb4 7.Rxb4 Nc5
8.Rh4 a6 (5.845.935) 781

15.01 0:13 +1.64 1.Kf1 f6 2.Be3 Qb5 3.b4 Rac8 4.Rxc8 Qxe2+
5.Kxe2 Rxc8 6.Bd2 Bxb4 7.Rxb4 Nc5
8.Rh4 a6 9.Rd4 (10.376.623) 797

16.01 0:24 +1.43 1.Kf1 f6 2.Be3 Qb5 3.b4 Rac8 4.Rxc8 Qxe2+
5.Kxe2 Rxc8 6.Bd2 Bxb4 7.Rxb4 Nc5
8.Rh4 h6 9.Rc4 Kf7 (19.744.946) 808

16.04 0:38 +1.62 1.Nd4 Nc5 2.b4 Ne4 3.Bf4 Rad8 4.Nxe6 fxe6
5.Qxe4 b5 6.Rc6 Bb6 7.Qxe6+ Rf7
8.Be3 Qb7 9.Bxb6 axb6 (31.114.274) 806

17.01 1:07 +1.80 1.Nd4 Nc5 2.b4 Ne4 3.Bf4 Rad8 4.Qc2 b5
5.Rc6 Bb6 6.Qxe4 Qxa2 7.Ne2 Rc8
8.Rxc8 Rxc8 9.Bd6 g6 (54.900.045) 808

17.03 3:28 +2.96 1.Bf6 Rfc8 2.Qe5 Rc5 3.Qg3 g6 4.Rxa4 Rac8
5.Qf4 Qd3 6.Rf1 Rf5 7.Qh4 Rh5 8.Qd4 Qb5
9.Be7 (166.570.436) 800

18.01 4:13 +2.86 1.Bf6 Rfc8 2.Qe5 Rc5 3.Qg3 g6 4.Rxa4 Rac8
5.Qf4 Qd3 6.Rf1 Rf5 7.Qh4 Rh5 8.Qd4 Qb5
9.h3 Rd5 (203.122.606) 801

19.01 6:25 +3.03 1.Bf6 Rfc8 2.Qe5 Rc5 3.Qg3 g6 4.Rxa4 Rac8
5.Qf4 Qd3 6.Rf1 Rf5 7.Qh4 Rh5 8.Qd4 Rd5
9.Qxd3 Rxd3 10.Rd4 Rxd4 11.Nxd4 (310.216.851) 803

20.01 17:24 +4.03 1.Bf6 Rfc8 2.Qe5 Rxc4 3.Qg5 Rg4
4.Qxg4 g6 5.Qxa4 b5 6.Qh4 Qc8 7.Rf1 Bd8
8.Ng5 h5 9.Qd4 Bxf6 10.Qxf6 Qd7
11.Rc1 (832.980.882) 797

best move: Bg5-f6 time: 19:32.296 min n/s: 800.566 nodes: 938.490.000

Very strange detail Tord, in Analysis mode the reported Hash saturation -I noticed it in this position but probably it happens in all testspositions-, all tests are being done in Shredder 9 UCI interface, was going down after a certain time and ended at Hash saturation: -321.-3% !

Result of the match against Belka is to follow later today I hope
Regards, Eelco