Banksia GUI released

[phhnguyen]

Hello Pham..

I have a request Could the timeout be switched off under tournament conditions?
Thank you for this super software.

Many greetings

The problem is that some engines may stop responding (but not crash technically). If BSG doesn't rule them by timeout, their games may stop forever. Look like we have no point to keep them running!

Note that we may play multi games concurrently over a long period thus it is not easy, and efficient to detect and solve them manually.

If you want to reduce timeout games, you may increase the value of the Margin when setting the Timer. When an engine is a timeout, BSG will wait for that margin period (give more chance for that engine to respond) before terminating that game.

Also, users could reset all timeout and/or crashed games with just a few clicks, easy to replay them.

[phhnguyen]

There is still a bug to fix about games.

I played a game between stockfish 10M nodes per move and stockfish1 100M nodes per move from the position after 1.g4 and it seems that the pgn is not correct.

pgn should start based on the fen 2...d5 and not 1.d5
I also do not understand why the score has only one digit after the first number when I believe stockfish show 2 moves after the digit and I can see 2 digits when I click on show scores and not copy pgn.

Thank you a lot for the report. I have been working to fix that bug.

[phhnguyen]

Question to ChatGPT and here is the answer:

Code: Select all

What is BanksiaGUI the chess GUI?

I am not familiar with a specific chess program called
"banksiagui". It is possible that it is a recently developed
chess software that uses a graphical user interface (GUI)
and has been named after the Banksia plant, which is
native to Australia.

It could also be a chess GUI that was developed by someone
or some company named Banksiagui.

Without more information, it is difficult to provide a specific
answer.

Look likes ChatGPT with data before 2021 doesn’t know about BanksiaGUI. It is a shame!

Save the answer here and wait for any improvement in coming time!

[Uri Blass]

Question to ChatGPT and here is the answer:

Code: Select all

What is BanksiaGUI the chess GUI?

I am not familiar with a specific chess program called
"banksiagui". It is possible that it is a recently developed
chess software that uses a graphical user interface (GUI)
and has been named after the Banksia plant, which is
native to Australia.

It could also be a chess GUI that was developed by someone
or some company named Banksiagui.

Without more information, it is difficult to provide a specific
answer.

Look likes ChatGPT with data before 2021 doesn’t know about BanksiaGUI. It is a shame!

Save the answer here and wait for any improvement in coming time!

This thread is from 2019 so data before 2021 should be enough for chatGPT to know about Banksia GUI

[Mark 1]

Thank you Pham

With margin it works .

Many greetings

[n4k3dw4ff13s]

I'm trying to create an opening book (polyglot) using Banksia Gui. My .pgn file is massive-- around 17.3 gb. Looking at the opening book creation setting in Book Tools, I'm not sure what some of the settings mean, and I don't see any documentation online:

min game length:30, min repeat:3, Add to: 20 plies, Score factors: wins: 5 draw: 0 loss:0 .

I'm sure some of the positions are duplicates (some with different move order, some with same move order.) I did not go through the .pgn with .pgn extract as I wanted to see how the Polyglot opening book file format and Banksia work. I'll try again later once I understand the settings and then again once I've pruned the large .pgn file.

Ideally, I'd want to store all the positions from every game, but is that unwise? What would be the pros and cons for storing all positions from every game in the .pgn file?

[Chessqueen]

I'm trying to create an opening book (polyglot) using Banksia Gui. My .pgn file is massive-- around 17.3 gb. Looking at the opening book creation setting in Book Tools, I'm not sure what some of the settings mean, and I don't see any documentation online:

min game length:30, min repeat:3, Add to: 20 plies, Score factors: wins: 5 draw: 0 loss:0 .

I'm sure some of the positions are duplicates (some with different move order, some with same move order.) I did not go through the .pgn with .pgn extract as I wanted to see how the Polyglot opening book file format and Banksia work. I'll try again later once I understand the settings and then again once I've pruned the large .pgn file.

Ideally, I'd want to store all the positions from every game, but is that unwise? What would be the pros and cons for storing all positions from every game in the .pgn file?

So far which is the best Opening Book under Banksia GUI available ?

[n4k3dw4ff13s]

Hello! I'm trying to create an opening book and I noticed something about how the program functions. It seems to load as many games from the .pgn into memory as possible (for me that's 12gb of games in memory) and it keeps 12gb of games in memory throughout the whole writing process. I wonder why this process isn't done in chunks. I noticed for an enterprise proprietary software like chessbase, they do this process in batches. I supposed they were thinking of everyday machines in mind. My .pgn file of 17.3 gb gets loaded into memory in chunks of 1.5-2.5gb at a time and only once the program writes those games to the disk (at 70-100mb/sec) does it load the next chunk.

It seems the games are skimmed in some manner at the start before this process begins allowing for this smooth clunking loading writing process.

I'm not sure about the speed differences in one loading method over the other, BUT I will say the chessbase opening book creation seems to have had user interaction in mind. I get the sense of progress when my fans turn on during the loading to memory phase and the disk writing phase and they quiet down as the next chunk is accessed and then back on again when the process repeats.

It's kinda not satisfying that everything is quiet as I wait for the final 38% percent of the progress bar to finish up during the disk writing phase.

I'd want to hear about why the process isn't done in chunks (not a critique but rather a curiosity) and if the structure of the .obs opening book is just better to be written to in this manner.

[phhnguyen]

So far which is the best Opening Book under Banksia GUI available ?

BSG supports books in Polyglot, PGN, EPD, and OOBS (Open Opening Book Standard, SQLite database). Users have plenty of books to choose from. I think "the best" is dependent much on purposes and personal tastes. I have read that many people talk about Cerebellum (Polyglot - comes already with BSG) and UHO (PGN).

If you or someone could, please suggest!

[phhnguyen]

Hello! I'm trying to create an opening book and I noticed something about how the program functions. It seems to load as many games from the .pgn into memory as possible (for me that's 12gb of games in memory) and it keeps 12gb of games in memory throughout the whole writing process. I wonder why this process isn't done in chunks. I noticed for an enterprise proprietary software like chessbase, they do this process in batches. I supposed they were thinking of everyday machines in mind. My .pgn file of 17.3 gb gets loaded into memory in chunks of 1.5-2.5gb at a time and only once the program writes those games to the disk (at 70-100mb/sec) does it load the next chunk.

It seems the games are skimmed in some manner at the start before this process begins allowing for this smooth clunking loading writing process.

I'm not sure about the speed differences in one loading method over the other, BUT I will say the chessbase opening book creation seems to have had user interaction in mind. I get the sense of progress when my fans turn on during the loading to memory phase and the disk writing phase and they quiet down as the next chunk is accessed and then back on again when the process repeats.

It's kinda not satisfying that everything is quiet as I wait for the final 38% percent of the progress bar to finish up during the disk writing phase.

I'd want to hear about why the process isn't done in chunks (not a critique but rather a curiosity) and if the structure of the .obs opening book is just better to be written to in this manner.

Actually, BSG processes PGN in chunks. It reads PGN block by block and then parses game by game. Later multi threads are applied to process games and build the opening tree. Processed games are dumped immediately to save space. I believe that makes BSG one of the fastest programs to build opening books.

(If you or someone needs/wants to look at that code, just look into OCGDB open source since BSG uses that code too for creating opening books).

Perhaps that is the problem with updating the progress bar. I will check it. Thanks.