Banksia GUI released

[Guenther]

The content looks like this:

Code: Select all

[Event "Book"]
[Site "RWBC"]
[Date "????.??.??"]
[Round "?"]
[White "ABC"]
[Black "XYZ"]
[Result "1/2-1/2"]

1. c4 e5 2. Nc3 Nc6 3. Nf3 Bb4 1/2-1/2

[Event "Book"]
[Site "RWBC"]
[Date "????.??.??"]
[Round "?"]
[White "ABC"]
[Black "XYZ"]
[Result "1/2-1/2"]

1. d4 c5 2. d5 d6 3. e4 g6 1/2-1/2

...

Thanks for the report.

"book" : null means there is no book, either they did not set or BSG can't read them. If you use an older version of BSG, there may be a bug (BSG tried to share books between normal games and tournaments but conflicts on writing thus books may be lost from configurations).

Can you try with the latest version (0.8)?
- Set up books, verify if they are saved down well on JSON files
- Play some games/tournament to see if those books work
- Check JSON files if any missing

Thanks again.

This was with version 0.8. It is not really a book, but pgn start positions (all 6 plies) as you can see above.
I set it in the tournament dialogue, but it never is remembered.
Is there another way to set a pgn file? The funny thing is it worked once for one game.

Edit:

[phhnguyen]

This was with version 0.8. It is not really a book, but pgn start positions (all 6 plies) as you can see above.
I set it in the tournament dialogue, but it never is remembered.
Is there another way to set a pgn file? The funny thing is it worked once for one game.

Edit:

Oops, I am sorry, it is a bug. BSG did not add new books correctly before creating tournaments (as well as saving back to JSON files). It will be fixed for the next release.

For the current version, you may fix by editing directly JSON file
- Stop BSG if it is running
- Open file "banksiatour.json" using any text editor
- Replace string "books" : null (we mentioned before) by bellow text:

Code: Select all

    "books" :
    [
      {
        "black" : true,
        "mode" : true,
        "name" : "3M_rand",
        "path" : "C:/Banksia/Openings/3M_rand.pgn",
        "type" : "pgn",
        "white" : true
      }
    ]

Take care if there is any redundant/missing comma.
- Save the file and open BSG.

Now you should see that book when creating a new tournament. The book should remain even you edit it too.

[Guenther]

This was with version 0.8. It is not really a book, but pgn start positions (all 6 plies) as you can see above.
I set it in the tournament dialogue, but it never is remembered.
Is there another way to set a pgn file? The funny thing is it worked once for one game.

Oops, I am sorry, it is a bug. BSG did not add new books correctly before creating tournaments (as well as saving back to JSON files). It will be fixed for the next release.

For the current version, you may fix by editing directly JSON file
- Stop BSG if it is running
- Open file "banksiatour.json" using any text editor
- Replace string "books" : null (we mentioned before) by bellow text:

Code: Select all

    "books" :
    [
      {
        "black" : true,
        "mode" : true,
        "name" : "3M_rand",
        "path" : "C:/Banksia/Openings/3M_rand.pgn",
        "type" : "pgn",
        "white" : true
      }
    ]

Take care if there is any redundant/missing comma.
- Save the file and open BSG.

Now you should see that book when creating a new tournament. The book should remain even you edit it too.

Thanks, that works now.

Meanwhile I found another serious bug, which obviously has to do with syzygy adudication.
Actually I hit the bug already 2 times by chance in the last 2 hours ;-)

When syzygy adjudication is enabled Banksia crashes, when the move which gives a syzygy postion (5 men in my case) is executed by an engine.
(windows crash popups)

Code: Select all

Problemsignatur:
  Problemereignisname:	APPCRASH
  Anwendungsname:	BanksiaGui.exe
  Anwendungsversion:	0.0.0.0
  Anwendungszeitstempel:	5ddc94e8
  Fehlermodulname:	BanksiaGui.exe
  Fehlermodulversion:	0.0.0.0
  Fehlermodulzeitstempel:	5ddc94e8
  Ausnahmecode:	c000001d
  Ausnahmeoffset:	0000000000040ade
  Betriebsystemversion:	6.1.7601.2.1.0.256.1
  Gebietsschema-ID:	1031
  Zusatzinformation 1:	1167
  Zusatzinformation 2:	1167b989b6d70415ffc3e1791f82dce4
  Zusatzinformation 3:	7ef4
  Zusatzinformation 4:	7ef49b073040fbbb436c3a7ef2366f6e

Example 1:
Here the move, which leads to 5 men syzgy is even hidden from the log - there is 90 seconds w/o any logging after bestmove g4f4,
but obviously Black had to play Rxf4, but a bit later Banksia crashed)

[d]8/8/8/6P1/5RrP/8/5k1K/8 b - - 0 98

Code: Select all

12:56:30 GreKo_90-64< position startpos moves d2d4 g8f6 b1c3 c7c5 d4c5 e7e6 e2e4 b8c6 c1e3 d8a5 f1d3 f8c5 e3c5 a5c5 g1f3 d7d6 e1g1 e6e5 a2a3 c8g4 h2h3 g4f3 d1f3 c6d4 f3d1 a8c8 c3e2 e8g8 b2b4 c5b6 e2d4 b6d4 f1e1 c8c3 d1e2 f8c8 e2e3 d4e3 e1e3 f6e8 b4b5 e8c7 e3e2 c7e6 e2d2 e6d4 a3a4 g8f8 a1a2 f8e7 a4a5 h7h5 a5a6 b7b6 a2b2 e7e6 g1f1 f7f6 f1e1 g7g6 e1f1 e6e7 f1g1 h5h4 g1h1 e7e6 h1h2 c8c7 h2g1 c7c5 g1h2 d6d5 f2f3 d5e4 f3e4 f6f5 d2f2 f5e4 d3e4 c5b5 b2b5 d4b5 e4g6 b5d6 f2f8 c3c7 f8h8 e6f6 g6h5 e5e4 h8h6 f6e5 h5g4 c7c2 h6h7 c2a2 h7a7 e5d5 g4d1 a2a3 a7a8 b6b5 a6a7 d5c6 d1g4 a3a1 g4e6 e4e3 e6g4 c6c5 g4f3 a1a2 a8h8 a2a7 h8h5 c5b4 h5h4 b4c3 h4h5 a7e7 f3e2 e7e6 h5d5 b5b4 e2g4 e6f6 g4e2 b4b3 d5d3 c3c2 d3e3 b3b2 e3g3 c2d2 e2d3 f6f1 d3g6 b2b1q g6b1 f1b1 g3f3 d2e2 h2g3 b1b7 g3h2 d6c4 f3f4 c4d2 f4f6 d2f1 h2g1 f1e3 g1h2 b7h7 f6f4 e3f1 h2g1 f1d2 f4f5 e2e3 g1h2 h7b7 g2g4 b7b2 h2g3 d2e4 g3h4 b2b8 g4g5 b8b4 f5f8 e4f6 h4g3 f6h5 g3g2 b4b2 g2h1 h5g3 h1g1 g3e2 g1f1 b2b1 f1g2 e2f4 g2g3 b1g1 g3h2 g1g2 h2h1 g2g3 h3h4 g3g4 h1h2 e3f2 f8f4
go wtime 87223 btime 76913 winc 0 binc 0 movestogo 23
12:56:30 GreKo_90-64> info multipv 1 depth 1 score cp 731 time 1 nodes 5 nps 5000 pv g4f4
...snip...
12:56:30 GreKo_90-64> info multipv 1 depth 20 score cp 731 time 7 nodes 132 nps 18000 pv g4f4
12:56:30 GreKo_90-64> info time 7 nodes 132 nps 18000
12:56:30 GreKo_90-64> info multipv 1 depth 21 score cp 746 time 355 nodes 765643 nps 2150000 pv g4f4 h2h3
12:56:30 GreKo_90-64> info time 356 nodes 765643 nps 2150000
12:56:33 GreKo_90-64> bestmove g4f4
12:58:03 AbsoluteZero_2440-64< isready
12:58:03 AbsoluteZero_2440-64> readyok
12:58:03 GreKo_90-64< isready
12:58:03 > Error: GreKo_90-64 stalled too long or crashed. Stopped!
=>nothing more

Example 2:
Here the log is stopped after the bestmove g6b1 by Black which also leads to syzygy 5 men on the board, but Banksia crashed.

[d]8/8/3n2B1/6kP/8/2K5/8/1n6 w - - 0 56

Code: Select all

16:01:34 GreKo_90-64> bestmove b2b1n
16:01:34 AbsoluteZero_2440-64< position startpos moves b2b3 d7d5 c1b2 c7c5 e2e3 a7a6 g1f3 g8f6 c2c4 e7e6 d2d4 c5d4 e3d4 b8c6 f1d3 c8d7 e1g1 c6b4 d3e2 b4c6 f3e5 f8d6 e5d7 d8d7 c4c5 d6f4 b1d2 e6e5 d2f3 e5d4 g2g3 f4c7 f3d4 c7e5 f2f4 e5d4 b2d4 c6d4 d1d4 e8g8 e2f3 d7c6 a1e1 f8c8 b3b4 a6a5 a2a3 a5b4 a3b4 c8d8 g3g4 f6e8 e1e7 e8c7 d4e5 d8d7 e7d7 c6d7 e5d6 a8d8 d6d7 d8d7 f1a1 g8f8 a1a7 c7b5 c5c6 b5a7 c6d7 f8e7 f3d5 b7b5 d5f7 a7c6 f7e8 e7d8 g4g5 c6b4 f4f5 b4c6 g1f1 b5b4 e8f7 d8d7 f7g8 d7d6 g8h7 b4b3 h2h4 b3b2 f5f6 g7f6 g5f6 c6d4 h7g6 d4f3 h4h5 f3e5 g6f5 e5f7 f1e2 d6e5 f5c2 e5f6 e2d3 f7d6 d3c3 f6g5 c2g6 b2b1n
go wtime 141105 btime 109736 winc 0 binc 0 movestogo 25
16:01:34 AbsoluteZero_2440-64> info depth 1 score cp 39 time 0 nodes 12 nps 244399 pv g6b1
16:01:34 AbsoluteZero_2440-64> info depth 2 score cp 39 time 0 nodes 40 nps 107382 pv g6b1 g5h5
...snip...
16:01:38 AbsoluteZero_2440-64> info depth 33 score cp -6 time 3410 nodes 6615577 nps 1939589 pv g6b1 g5h5 b1d3 h5g5 c3d4 g5f6 d4d5 d6b5 d5e4 b5c3 e4d4 c3b5 d4c5 b5d4 c5b4 f6e5 b4c4 d4f5 d3e2 f5d6 c4d3 e5f4 e2h5 f4e5 h5f3 d6b5 d3d2 e5f6 d2e1 f6e5 e1f2 b5d6 f2e3 d6c4 e3e2 e5d4 e2f2
16:01:38 AbsoluteZero_2440-64> bestmove g6b1
=>nothing more

[Gabor Szots]

Version 0.8 released

As usual, it fixed some bugs. The most serious one it could not detect some options which words in rare orders; a bug about not showing pv for engines pondering; table's headers become black in Linux.

New features:

• Shuffle tournament games Carousel

• Customizable color for highlighting boards in Playing boards dialog and for move highlights

• Tournament result tables are shown in different tabs (not all in one)

Thanks for carousel mode, Nguyen. Maybe it can be further improved by making the gauntlet engine play the first round with white, the second with black, etc., repeating.

BTW, what is 'Mode' at the adjudication section?

[phhnguyen]

This was with version 0.8. It is not really a book, but pgn start positions (all 6 plies) as you can see above.
I set it in the tournament dialogue, but it never is remembered.
Is there another way to set a pgn file? The funny thing is it worked once for one game.

Oops, I am sorry, it is a bug. BSG did not add new books correctly before creating tournaments (as well as saving back to JSON files). It will be fixed for the next release.

For the current version, you may fix by editing directly JSON file
- Stop BSG if it is running
- Open file "banksiatour.json" using any text editor
- Replace string "books" : null (we mentioned before) by bellow text:

Code: Select all

    "books" :
    [
      {
        "black" : true,
        "mode" : true,
        "name" : "3M_rand",
        "path" : "C:/Banksia/Openings/3M_rand.pgn",
        "type" : "pgn",
        "white" : true
      }
    ]

Take care if there is any redundant/missing comma.
- Save the file and open BSG.

Now you should see that book when creating a new tournament. The book should remain even you edit it too.

Thanks, that works now.

Meanwhile I found another serious bug, which obviously has to do with syzygy adudication.
Actually I hit the bug already 2 times by chance in the last 2 hours

When syzygy adjudication is enabled Banksia crashes, when the move which gives a syzygy postion (5 men in my case) is executed by an engine.
(windows crash popups)

Code: Select all

Problemsignatur:
  Problemereignisname:	APPCRASH
  Anwendungsname:	BanksiaGui.exe
  Anwendungsversion:	0.0.0.0
  Anwendungszeitstempel:	5ddc94e8
  Fehlermodulname:	BanksiaGui.exe
  Fehlermodulversion:	0.0.0.0
  Fehlermodulzeitstempel:	5ddc94e8
  Ausnahmecode:	c000001d
  Ausnahmeoffset:	0000000000040ade
  Betriebsystemversion:	6.1.7601.2.1.0.256.1
  Gebietsschema-ID:	1031
  Zusatzinformation 1:	1167
  Zusatzinformation 2:	1167b989b6d70415ffc3e1791f82dce4
  Zusatzinformation 3:	7ef4
  Zusatzinformation 4:	7ef49b073040fbbb436c3a7ef2366f6e

Example 1:
Here the move, which leads to 5 men syzgy is even hidden from the log - there is 90 seconds w/o any logging after bestmove g4f4,
but obviously Black had to play Rxf4, but a bit later Banksia crashed)

[d]8/8/8/6P1/5RrP/8/5k1K/8 b - - 0 98

Code: Select all

12:56:30 GreKo_90-64< position startpos moves d2d4 g8f6 b1c3 c7c5 d4c5 e7e6 e2e4 b8c6 c1e3 d8a5 f1d3 f8c5 e3c5 a5c5 g1f3 d7d6 e1g1 e6e5 a2a3 c8g4 h2h3 g4f3 d1f3 c6d4 f3d1 a8c8 c3e2 e8g8 b2b4 c5b6 e2d4 b6d4 f1e1 c8c3 d1e2 f8c8 e2e3 d4e3 e1e3 f6e8 b4b5 e8c7 e3e2 c7e6 e2d2 e6d4 a3a4 g8f8 a1a2 f8e7 a4a5 h7h5 a5a6 b7b6 a2b2 e7e6 g1f1 f7f6 f1e1 g7g6 e1f1 e6e7 f1g1 h5h4 g1h1 e7e6 h1h2 c8c7 h2g1 c7c5 g1h2 d6d5 f2f3 d5e4 f3e4 f6f5 d2f2 f5e4 d3e4 c5b5 b2b5 d4b5 e4g6 b5d6 f2f8 c3c7 f8h8 e6f6 g6h5 e5e4 h8h6 f6e5 h5g4 c7c2 h6h7 c2a2 h7a7 e5d5 g4d1 a2a3 a7a8 b6b5 a6a7 d5c6 d1g4 a3a1 g4e6 e4e3 e6g4 c6c5 g4f3 a1a2 a8h8 a2a7 h8h5 c5b4 h5h4 b4c3 h4h5 a7e7 f3e2 e7e6 h5d5 b5b4 e2g4 e6f6 g4e2 b4b3 d5d3 c3c2 d3e3 b3b2 e3g3 c2d2 e2d3 f6f1 d3g6 b2b1q g6b1 f1b1 g3f3 d2e2 h2g3 b1b7 g3h2 d6c4 f3f4 c4d2 f4f6 d2f1 h2g1 f1e3 g1h2 b7h7 f6f4 e3f1 h2g1 f1d2 f4f5 e2e3 g1h2 h7b7 g2g4 b7b2 h2g3 d2e4 g3h4 b2b8 g4g5 b8b4 f5f8 e4f6 h4g3 f6h5 g3g2 b4b2 g2h1 h5g3 h1g1 g3e2 g1f1 b2b1 f1g2 e2f4 g2g3 b1g1 g3h2 g1g2 h2h1 g2g3 h3h4 g3g4 h1h2 e3f2 f8f4
go wtime 87223 btime 76913 winc 0 binc 0 movestogo 23
12:56:30 GreKo_90-64> info multipv 1 depth 1 score cp 731 time 1 nodes 5 nps 5000 pv g4f4
...snip...
12:56:30 GreKo_90-64> info multipv 1 depth 20 score cp 731 time 7 nodes 132 nps 18000 pv g4f4
12:56:30 GreKo_90-64> info time 7 nodes 132 nps 18000
12:56:30 GreKo_90-64> info multipv 1 depth 21 score cp 746 time 355 nodes 765643 nps 2150000 pv g4f4 h2h3
12:56:30 GreKo_90-64> info time 356 nodes 765643 nps 2150000
12:56:33 GreKo_90-64> bestmove g4f4
12:58:03 AbsoluteZero_2440-64< isready
12:58:03 AbsoluteZero_2440-64> readyok
12:58:03 GreKo_90-64< isready
12:58:03 > Error: GreKo_90-64 stalled too long or crashed. Stopped!
=>nothing more

Example 2:
Here the log is stopped after the bestmove g6b1 by Black which also leads to syzygy 5 men on the board, but Banksia crashed.

[d]8/8/3n2B1/6kP/8/2K5/8/1n6 w - - 0 56

Code: Select all

16:01:34 GreKo_90-64> bestmove b2b1n
16:01:34 AbsoluteZero_2440-64< position startpos moves b2b3 d7d5 c1b2 c7c5 e2e3 a7a6 g1f3 g8f6 c2c4 e7e6 d2d4 c5d4 e3d4 b8c6 f1d3 c8d7 e1g1 c6b4 d3e2 b4c6 f3e5 f8d6 e5d7 d8d7 c4c5 d6f4 b1d2 e6e5 d2f3 e5d4 g2g3 f4c7 f3d4 c7e5 f2f4 e5d4 b2d4 c6d4 d1d4 e8g8 e2f3 d7c6 a1e1 f8c8 b3b4 a6a5 a2a3 a5b4 a3b4 c8d8 g3g4 f6e8 e1e7 e8c7 d4e5 d8d7 e7d7 c6d7 e5d6 a8d8 d6d7 d8d7 f1a1 g8f8 a1a7 c7b5 c5c6 b5a7 c6d7 f8e7 f3d5 b7b5 d5f7 a7c6 f7e8 e7d8 g4g5 c6b4 f4f5 b4c6 g1f1 b5b4 e8f7 d8d7 f7g8 d7d6 g8h7 b4b3 h2h4 b3b2 f5f6 g7f6 g5f6 c6d4 h7g6 d4f3 h4h5 f3e5 g6f5 e5f7 f1e2 d6e5 f5c2 e5f6 e2d3 f7d6 d3c3 f6g5 c2g6 b2b1n
go wtime 141105 btime 109736 winc 0 binc 0 movestogo 25
16:01:34 AbsoluteZero_2440-64> info depth 1 score cp 39 time 0 nodes 12 nps 244399 pv g6b1
16:01:34 AbsoluteZero_2440-64> info depth 2 score cp 39 time 0 nodes 40 nps 107382 pv g6b1 g5h5
...snip...
16:01:38 AbsoluteZero_2440-64> info depth 33 score cp -6 time 3410 nodes 6615577 nps 1939589 pv g6b1 g5h5 b1d3 h5g5 c3d4 g5f6 d4d5 d6b5 d5e4 b5c3 e4d4 c3b5 d4c5 b5d4 c5b4 f6e5 b4c4 d4f5 d3e2 f5d6 c4d3 e5f4 e2h5 f4e5 h5f3 d6b5 d3d2 e5f6 d2e1 f6e5 e1f2 b5d6 f2e3 d6c4 e3e2 e5d4 e2f2
16:01:38 AbsoluteZero_2440-64> bestmove g6b1
=>nothing more

Thanks a lot for the report. I have been working on that issue (a bit hard for me to reproduce)

[phhnguyen]

Thanks for carousel mode, Nguyen. Maybe it can be further improved by making the gauntlet engine play the first round with white, the second with black, etc., repeating.

Thanks a lot for the idea. It has implemented for the next release.

BTW, what is 'Mode' at the adjudication section?

Simply it is "turn" or "switch" to turn on, off

[Guenther]

Thanks a lot for the report. I have been working on that issue (a bit hard for me to reproduce)

Do you have a debug version? I tried it with other more known engines and the result was always the same.
(Banksia crashing when a move would come, which would give a 5 men syzygy position on the board)

More info:
Win7-64 old quadcore w/o popcount and SSE 4.1. (May be there is still a little incompatibility with Win7 despite the dlls?)
Best would be someone with Win7 would reproduce the problem too...
Full 3-5 men Syzygy (290 files) in one folder set as directory in Banksia and set for adjudication.

BTW I don't understand the count (435) in Banksia?

in general settings


in tournament settings

[phhnguyen]

Thanks a lot for the report. I have been working on that issue (a bit hard for me to reproduce)

Do you have a debug version? I tried it with other more known engines and the result was always the same.
(Banksia crashing when a move would come, which would give a 5 men syzygy position on the board)

More info:
Win7-64 old quadcore w/o popcount and SSE 4.1.
Full 3-5 men Syzygy (290 files) in one folder set as directory in Banksia and set for adjudication.

BTW I don't understand the count (435) in Banksia?

in general settings


in tournament settings

I will send you a debug version about tomorrow (fixing few things first). Thanks a lot for your help.

435 is the number of "good" Syzygy endgames. The table lists all endgames BSG found and could read. Those endgames are counted from 1. If you scroll down to the end of the list the last count number must be 435.

The idea behinds that information is to help users quickly check their Syzygy endgames if they are readable, there is any missing/corrupted. It is not easy task for humans when the total of files is large (from my experience). I will show missing/corrupted files in later versions.

Edited: Oops, I understood your question now. So you mean why the number is 435 but not 290?

I don't know. In my computer, all 5 men is 288, closed to 290 and reasonable. Perhaps you need to scroll down to see any redundant / 6 men?

[Guenther]

BTW I don't understand the count (435) in Banksia?

I will send you a debug version about tomorrow (fixing few things first). Thanks a lot for your help.

435 is the number of "good" Syzygy endgames. The table lists all endgames BSG found and could read. Those endgames are counted from 1. If you scroll down to the end of the list the last count number must be 435.

The idea behinds that information is to help users quickly check their Syzygy endgames if they are readable, there is any missing/corrupted. It is not easy task for humans when the total of files is large (from my experience). I will show missing/corrupted files in later versions.

Edit: Oops, I understood your question now. So you mean why the number is 435 but not 290?

I don't know. In my computer, all my 5 men is 288, reasonable. Perhaps you need to scroll down to see any redundant?

No , all 3,4,5 men dtz and wdl are 290 together. It seems to me now you count one set twice?
145+145 = 290 + 1/2*290 = 435

[phhnguyen]

No , all 3,4,5 men dtz and wdl are 290 together. It seems to me now you count one set twice?
145+145 = 290 + 1/2*290 = 435

Thanks for the hint. Let me check and I will be back on this issue. BTW, I am going to add some code thus users can detect any issues with Syzygy files faster.