BBChess 1.2a : 2287

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

User avatar
Le Fou numerique
Posts: 859
Joined: Wed Apr 05, 2006 9:58 pm
Location: Elsass

BBChess 1.2a : 2287

Post by Le Fou numerique »

Hi,

ImageBBChess 1.2a first rating in the UEL: 2287.

PGN on the Division Minime page.

Code: Select all

Tournoi d'entrée HS-Book.abk
Le Fou numerique, 2007.10.09 - 2007.10.10
                              Score       1    2 
--------------------------------------------------
 1: BBChess 1.2a            48.5 / 126   49w- 49b+
 2: Aice 0.99.2              2.0 /  2    1b+  1w+
 3: Alaric 707               2.0 /  2    1b+  1w+
 4: Arasan 10.0              2.0 /  2    1b+  1w+
 5: Aristarch 4.50           2.0 /  2    1b+  1w+
 6: Booot 4.13.1             2.0 /  2    1b+  1w+
 7: Colossus 2007d           2.0 /  2    1b+  1w+
 8: Flux 2.2                 2.0 /  2    1b+  1w+
 9: Fruit 2.3.1              2.0 /  2    1b+  1w+
10: Glaurung 2 - epsilon5    2.0 /  2    1b+  1w+
11: Homer 2.01               2.0 /  2    1b+  1w+
12: Ktulu 4.2                2.0 /  2    1b+  1w+
13: Lime 62                  2.0 /  2    1b+  1w+
14: List 5.12                2.0 /  2    1b+  1w+
15: Movei 0.08.438           2.0 /  2    1b+  1w+
16: Pharaon 3.5.1            2.0 /  2    1b+  1w+
17: Ruffian 1.0.5            2.0 /  2    1b+  1w+
18: Yace Paderborn           2.0 /  2    1b+  1w+
19: Arion 1.7                1.5 /  2    1b+  1w=
20: Mediocre 0.332           1.5 /  2    1b=  1w+
21: Hamsters 0.5             1.5 /  2    1b+  1w=
22: Naum 2.0                 1.5 /  2    1b+  1w=
23: Patzer 3.80              1.5 /  2    1b+  1w=
24: Dragon 4.6               1.5 /  2    1b=  1w+
25: Pupsi 0.19               1.5 /  2    1b+  1w=
26: AnMon 5.60               1.5 /  2    1b=  1w+
27: Sjeng 12.13              1.5 /  2    1b+  1w=
28: SlowChess Blitz WV2.1    1.5 /  2    1b=  1w+
29: Snitch 1.6.2             1.5 /  2    1b+  1w=
30: Spike 1.2 Turin          1.5 /  2    1b+  1w=
31: SOS 5.1                  1.5 /  2    1b=  1w+
32: Tao 5.6                  1.5 /  2    1b=  1w+
33: Twisted Logic 20070915   1.5 /  2    1b+  1w=
34: Ufim 8.02                1.5 /  2    1b=  1w+
35: Abrok 5.0                1.5 /  2    1b=  1w+
36: Adam 3.1                 1.0 /  2    1b-  1w+
37: Chispa 4.0.3             1.0 /  2    1b=  1w=
38: Rotor 0.2                1.0 /  2    1b-  1w+
39: Alfil 7.6                1.0 /  2    1b+  1w-
40: Delfi 5.2                1.0 /  2    1b-  1w+
41: Diablo 0.5.1             1.0 /  2    1b-  1w+
42: Anechka 0.08             1.0 /  2    1b-  1w+
43: Madeleine 0.2            1.0 /  2    1b-  1w+
44: Feuerstein 0.4.5.2       1.0 /  2    1b+  1w-
45: Amyan 1.597              1.0 /  2    1b+  1w-
46: Muse 0.899b              1.0 /  2    1b+  1w-
47: Asterisk 0.6             1.0 /  2    1b=  1w=
48: Gaia 3.5                 1.0 /  2    1b+  1w-
49: Zappa 1.1                1.0 /  2    1b+  1w-
50: Atlas 2.20               0.5 /  2    1b-  1w=
51: Delphil 1.8              0.5 /  2    1b=  1w-
52: Hermann 2.0              0.5 /  2    1b=  1w-
53: GreKo 5.4                0.5 /  2    1b-  1w=
54: Nejmet 3.07              0.5 /  2    1b=  1w-
55: WJChess v1.64            0.5 /  2    1b-  1w=
56: Matheus 2.3              0.5 /  2    1b=  1w-
57: Pepito v1.59             0.5 /  2    1b-  1w=
58: Queen 3.09               0.0 /  2    1b-  1w-
59: Gibbon 2.42c             0.0 /  2    1b-  1w-
60: Ayito 0.2.994            0.0 /  2    1b-  1w-
61: AliChess 4.08            0.0 /  2    1b-  1w-
62: Monarch 1.7              0.0 /  2    1b-  1w-
63: Counter 0.6              0.0 /  2    1b-  1w-
64: LittleThought 1.01       0.0 /  2    1b-  1w-
--------------------------------------------------
126 games: +48 =29 -49
Regards,
Patrick
Top
User avatar
Graham Banks
Posts: 44595
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: BBChess 1.2a : 2287

Post by Graham Banks »

Hi Patrick,

should the "fixed parallel search" difference between BBChess 1.2 and BBChess 1.2a make any difference to strength on a single cpu?
I'm no programmer - that's why I ask.

Regards, Graham.
gbanksnz at gmail.com
Top
User avatar
pedrox
Posts: 1056
Joined: Fri Mar 10, 2006 6:07 am
Location: Basque Country (Spain)

Re: BBChess 1.2a : 2287

Post by pedrox »

No

The program can follow 2 ways, one processor or several, if the error were in the search in parallel it would not have to affect in a single processor.

Different it is if the author made more changes.
Top
User avatar
Le Fou numerique
Posts: 859
Joined: Wed Apr 05, 2006 9:58 pm
Location: Elsass

Re: BBChess 1.2a : 2287

Post by Le Fou numerique »

Graham Banks wrote:Hi Patrick,

should the "fixed parallel search" difference between BBChess 1.2 and BBChess 1.2a make any difference to strength on a single cpu?
I'm no programmer - that's why I ask.

Regards, Graham.
Hi Graham,

It seems the results between the two versions are the same 2287 to 2292.

Regards,
Patrick
Top
User avatar
Graham Banks
Posts: 44595
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: BBChess 1.2a : 2287

Post by Graham Banks »

pedrox wrote:No

The program can follow 2 ways, one processor or several, if the error were in the search in parallel it would not have to affect in a single processor.

Different it is if the author made more changes.
That's what I thought. As the fix to the parallel search bugs was the only change from BBChess 1.2, I think I'm safe to count 1.2a as exactly the same engine on a single cpu.

Thanks, Graham.
gbanksnz at gmail.com
Top
User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 8:33 pm
Location: Antalya, Turkey

Re: BBChess 1.2a : 2287

Post by beachknight »

Hello Graham,

If you look at both source codes, there could be changes in eval.c and
other routines as well.

I took a random position for a test.

Code: Select all


FEN: 1r1rb1k1/5ppp/4p3/1p1p3P/1q2P2Q/pN3P2/PPP4P/1K1R2R1 w - - 0 1 

BBChess 1.2a:
  5/21	00:00,436	      96.214	279.895	+0,99	1.Qf6 g6 2.Rd4 Qd6 3.hxg6
  6/21	00:00,656	     165.103	285.583	+0,98	1.Qf6 Qf8 2.h6 g6
  6/21	00:00,671	     195.123	297.330	+0,94	1.Qf6 Qf8 2.h6 g6
  6/21	00:00,874	     271.986	310.841	+0,70	1.Qf6 Qf8 2.h6 g6
  6/21	00:01,092	     353.493	323.193	+0,41	1.Qf6 Qf8 2.bxa3 Kh8 3.Qh4 Qxa3
  6/21	00:01,155	     380.454	329.041	+0,42	1.h6 g6 2.bxa3
  6/21	00:01,312	     399.527	327.817	+0,46	1.h6 g6 2.Rd4 Qd6 3.bxa3
  6/21	00:01,312	     419.815	331.705	+0,54	1.h6 g6 2.Rd4 Qd6 3.bxa3
  6/21	00:01,342	     452.886	337.031	+0,70	1.h6 g6 2.Rd4 Qd6 3.exd5
  6/21	00:01,437	     497.458	346.057	+0,74	1.h6 g6 2.Rd4 Qf8 3.exd5 axb2 4.dxe6
  7/21	00:02,467	     906.286	367.103	+0,74	1.h6 g6 2.Rd4 Qf8 3.exd5 axb2 4.dxe6
  8/23	00:03,218	   1.177.063	365.689	+0,73	1.h6 g6 2.Rd4 Qd6 3.exd5 exd5
  8/23	00:03,578	   1.322.308	369.553	+0,70	1.h6 g6 2.Rd4 Qd6 3.exd5 exd5 4.Rf4 axb2 5.Kxb2
  8/23	00:06,936	   2.700.698	389.289	+0,71	1.Rd4
  8/23	00:07,093	   2.756.751	388.616	+0,74	1.Rd4 Qf8 2.h6 g6 3.exd5 axb2 4.dxe6 Rxd4 5.Nxd4
  9/23	00:07,609	   2.991.577	393.143	+0,73	1.Rd4 Qf8 2.h6 g6 3.exd5 Rxd5 4.Rxd5 exd5
  9/23	00:08,140	   3.237.760	398.493	+0,69	1.Rd4 Qf8 2.h6 g6 3.exd5 Rxd5 4.Rxd5 exd5
  9/23	00:08,702	   3.506.381	402.887	+0,61	1.Rd4 Qf8 2.h6 g6 3.exd5 Rxd5 4.Rxd5 exd5
  9/23	00:09,562	   3.893.925	407.207	+0,47	1.Rd4 Qf8 2.h6 g6 3.exd5 Rxd5 4.Rxd5 exd5 5.Qf4
  9/23	00:11,015	   4.488.964	407.508	+0,48	1.Qf6 Qf8 2.bxa3 Kh8 3.Qh4 Qxa3
  9/23	00:11,686	   4.769.660	408.099	+0,49	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6 5.Qe5
 10/24	00:21,499	   8.907.048	414.281	+0,50	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6 5.Qe5
 10/24	00:22,217	   9.193.014	413.750	+0,51	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6 5.Qe5 Rc8
 11/26	00:29,468	  12.186.418	413.536	+0,50	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6 5.Qe5 Rc8
 11/26	00:34,202	  14.203.193	415.260	+0,46	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6 5.Qe5 Rc8
 11/26	00:39,484	  16.448.509	416.582	+0,44	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 Bd7 5.Qe5 Rd8 6.Qxd5
 11/26	00:40,734	  17.022.319	417.885	+0,45	1.Rd4 Qf8 2.h6 g6 3.exd5 Rxd5 4.Rxd5 exd5 5.Qf4
 11/26	00:45,249	  18.894.155	417.550	+0,45	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 axb2 5.Nf5 g6 6.Kxb2 Ra8
 12/28	01:46,234	  44.833.633	422.025	+0,44	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 axb2 5.Nf5 Bd7
 12/28	01:57,422	  49.636.727	422.777	+0,40	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 axb2 5.Nf5 Bd7
 12/29	02:12,203	  55.879.282	422.677	+0,38	1.Rd4 Qf8 2.e5 g6 3.hxg6 hxg6 4.Rf4 Qg7 5.Rf6 axb2 6.Kxb2 Rbc8 7.Nd4
 12/29	02:17,218	  57.982.290	422.553	+0,39	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 Bd7 5.Qe5 Rd8 6.Qxd5
 12/29	02:18,655	  58.566.423	422.385	+0,43	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 Bd7 5.Qe5 Rd8
 12/29	02:23,531	  60.666.463	422.670	+0,46	1.Qf6 Qf8 2.exd5 exd5 3.bxa3 Kh8 4.Qe5 f6 5.Qf5 Qxa3 6.Rxd5
 13/29	02:59,577	  75.505.955	420.462	+0,46	1.Qf6 Qf8 2.exd5 exd5 3.bxa3 Ra8
 14/34	07:12,264	 182.775.920	422.847	+0,47	1.Qf6 Qf8 2.exd5 exd5 3.bxa3 Ra8 4.Rde1 Rdc8 5.Nd4
 14/34	07:30,047	 190.493.502	423.274	+0,47	1.Qf6 Qf8 2.exd5 exd5 3.bxa3 Kh8 4.Qh4 f6 5.f4
 15/34	10:42,687	 275.832.176	429.185	+0,46	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 Bd7 5.Qe5 Re8
 15/34	11:51,983	 306.594.117	430.619	+0,42	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 Bd7 5.Qe5 Re8
 15/34	13:34,156	 351.320.334	431.514	+0,34	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 Bd7 5.Qe5 Re8
 15/34	15:24,342	 399.059.707	431.722	+0,34	1.Qf6 Qf8 2.bxa3 Rbc8 3.exd5 Rxd5 4.Rxd5 exd5 5.Qe5 Bc6 6.Nd4
 15/37	43:23,889	1.137.807.013	436.964	+0,35	1.Rg4
 15/37	53:25,577	1.396.517.483	435.652	+0,35	1.Rg4 g6 2.exd5 Qf8 3.Rgg1 axb2 4.hxg6 hxg6 5.dxe6
Best move: Rg4, Value: +0,35, Depth: 16/37, Time: 1:13:07,308, 1.854.394.432 Nodes, 438.295 N/sec.



FEN: 1r1rb1k1/5ppp/4p3/1p1p3P/1q2P2Q/pN3P2/PPP4P/1K1R2R1 w - - 0 1 

BBChess 1.2:
  5/21	00:00,328	      80.309	256.988	+0,99	1.Qf6 g6 2.Rd4 Qd6 3.hxg6
  6/21	00:00,578	     141.072	265.547	+0,98	1.Qf6 Qf8 2.h6 g6
  6/21	00:00,611	     163.347	275.110	+0,94	1.Qf6 Qf8 2.h6 g6
  6/21	00:00,766	     215.770	287.693	+0,70	1.Qf6 Qf8 2.h6 g6
  6/21	00:00,922	     270.364	298.332	+0,41	1.Qf6 Qf8 2.bxa3 Kh8 3.Qh4 Qxa3
  6/21	00:00,985	     295.659	305.196	+0,42	1.h6 g6 2.bxa3
  6/21	00:01,079	     314.644	305.109	+0,46	1.h6 g6 2.Rd4 Qd6 3.bxa3
  6/21	00:01,094	     334.898	310.630	+0,54	1.h6 g6 2.Rd4 Qd6 3.bxa3
  6/21	00:01,172	     360.754	316.277	+0,70	1.h6 g6 2.Rd4 Qd6 3.exd5
  6/21	00:01,267	     403.512	322.809	+0,74	1.h6 g6 2.Rd4 Qf8 3.exd5 axb2 4.dxe6
  7/21	00:02,219	     755.832	343.072	+0,74	1.h6 g6 2.Rd4 Qf8 3.exd5 axb2 4.dxe6
  8/23	00:02,860	     961.812	338.219	+0,73	1.h6 g6 2.Rd4 Qd6 3.exd5 exd5
  8/23	00:03,188	   1.089.258	343.411	+0,70	1.h6 g6 2.Rd4 Qd6 3.exd5 exd5 4.Rf4 axb2 5.Kxb2
  8/23	00:05,063	   1.807.751	358.192	+0,71	1.Rd4
  8/23	00:05,158	   1.837.141	357.376	+0,74	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6
  9/23	00:06,578	   2.403.000	366.171	+0,73	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6
  9/23	00:07,064	   2.612.867	370.783	+0,69	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6
  9/23	00:07,595	   2.839.255	374.664	+0,61	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6
  9/23	00:08,453	   3.186.778	377.692	+0,47	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6 5.Qf4
  9/23	00:09,923	   3.741.382	377.678	+0,48	1.Qf6 Qf8 2.bxa3 Kh8 3.Qh4 Qxa3
  9/23	00:10,564	   3.980.735	377.432	+0,49	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6 5.Qe5
 10/27	00:22,470	   8.580.376	382.146	+0,48	1.Qf6 Qf8 2.exd5 exd5 3.Nd4 Bd7 4.h6 g6
 10/27	00:22,641	   8.644.574	382.080	+0,48	1.Qf6 Qf8 2.exd5 exd5 3.bxa3 Kh8 4.Qf5 Bc6 5.h6
 11/27	00:40,939	  15.536.407	379.660	+0,49	1.Qf6 Qf8 2.exd5 exd5 3.bxa3 Kh8 4.Qf5 Bc6 5.h6
 11/27	00:43,626	  16.577.807	380.143	+0,49	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Qe5 Rc8 5.h6 g6
 11/27	00:44,517	  16.934.638	380.553	+0,50	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6 5.Qf4
 11/27	00:47,750	  18.177.394	380.927	+0,50	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 axb2 5.Nf5 g6 6.Kxb2
 12/27	00:57,423	  21.942.849	382.342	+0,49	1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 axb2 5.Nf5 Bd7
 12/28	01:10,423	  27.065.914	384.424	+0,46	1.Rd4 Qf8 2.e5 g6 3.Rf4 Qg7 4.Rf6 axb2 5.hxg6 hxg6 6.Kxb2 b4
 13/31	02:54,469	  68.090.256	390.306	+0,46	1.Rd4 Qf8 2.e5 g6 3.Rf4 Qg7 4.Rf6 axb2 5.Nd4 Qh6 6.hxg6 Qxh4 7.gxf7+
 14/31	04:33,892	 105.371.940	384.744	+0,45	1.Rd4 Qd6 2.exd5 exd5 3.bxa3 Ra8 4.Rdg4 g6
 14/31	05:47,736	 134.760.841	387.557	+0,44	1.Rd4 Qd6 2.Qf6 Qf8 3.exd5 Rxd5 4.Rxd5 exd5 5.Qe5 Rc8 6.h6 g6
 14/31	06:18,970	 147.205.241	388.452	+0,45	1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Qe5 Rc8 5.h6 g6
 14/31	06:49,375	 159.101.006	388.658	+0,48	1.Qf6 Qf8 2.exd5 exd5 3.bxa3 Kh8 4.Qb2 f6 5.h6 g6
 15/36	16:02,954	 381.352.395	396.030	+0,47	1.Qf6 Qf8 2.exd5 exd5 3.bxa3 Kh8 4.Qb2 f6 5.h6 g6
Best move: Qf6, Value: +0,47, Depth: 15/36, Time: 16:08,699, 383.726.939 Nodes, 396.149 N/sec.
Up to 10 % speedup in knps figures for a computer with 2 threads.
hi, merhaba, hallo HT
Top
User avatar
Graham Banks
Posts: 44595
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: BBChess 1.2a : 2287

Post by Graham Banks »

[quote="beachknight"]Hello Graham,

If you look at both source codes, there could be changes in eval.c and
other routines as well.quote]

Okay - thanks Harun. I guess I'd best contact the engine author to be safe then.

Regards, Graham.
gbanksnz at gmail.com
Top
User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 8:33 pm
Location: Antalya, Turkey

Re: BBChess 1.2a : 2287

Post by beachknight »

Okay - thanks Harun. I guess I'd best contact the engine author to be safe then.
Hello Graham,

Another way:

Just take the position given above.

Run 15 mins with both engines.

And post the results here.

Lets wait if we see any difference.

Best,
hi, merhaba, hallo HT
Top
borko

Re: BBChess 1.2a : 2287

Post by borko »

BBChess 1.2 had a problem with transposition table in parallel search. In version 1.2a this bug was fixed and engine become more stable. Changes in transposition table influence to single thread search. Engine become a little bit slower in single thread search but faster in multi thread search and more stable.
Top
User avatar
Graham Banks
Posts: 44595
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: BBChess 1.2a : 2287

Post by Graham Banks »

borko wrote:BBChess 1.2 had a problem with transposition table in parallel search. In version 1.2a this bug was fixed and engine become more stable. Changes in transposition table influence to single thread search. Engine become a little bit slower in single thread search but faster in multi thread search and more stable.
Okay - thanks for explaining that Borko.
Looks like I can discard my BBChess 1.2 games and concentrate on BBChess 1.2a.

Regards, Graham.
gbanksnz at gmail.com
Top

Return to “Computer Chess Club: Tournaments and Matches”