Winboard Forum

ChessGUI 0.124 has just been "upped".

I have not tested it, but this version should be portable.
For portability :
- first install ChessGUI, engines and tablebases all in one hd partition.
- use uci engines which, apart from their own folders, only use the NalimovPath got from GUI.
- use wb engines which use only their own folders.
- portable engines must be marked in ChessGUI as portable.
- installed ChessGUI, engines and tablebases can then be copied to stick
- ChessGUI ships with a full set of 3-men tablebases
- on remote computer, tell ChessGUI in Preferences to work in portable mode and set NalimovPath only in GUI Preferences.

If it works for anybody, please report.

I have seen ChessGUI 0.123 and ChessGUI 0.124 exit unexpectedly in the interval between games. Maybe it affects only swiss tournaments ?
Watch out and report !

Matthias.

hi Matthias,

nice job on ChessGUI, it is quite handy.

On ChessGU 0.123, I've had some problems with the engine list, though: sometimes the list was completely erased, other, duplicates of all or almost all engines were created, e.g.

Alaric 707 :: Flags = UCI2 Can_Analyze Remis_Book
Alaric707 :: Flags = No_GUI_Book

the first is the one created by me, the second appeared after a restart of ChessGUI.

Also, I think Glaurung had problems, I had to install more than once, when trying to set up sometimes ChessGui would query and Glaurung not respond, and finally, the first tournament Glaurung lost all its games, but blundering, not being beaten normally.

I can try to provide more information, if you need.

anyway, thanks, nice gui.
horacio

nepossiver wrote:hi Matthias,

nice job on ChessGUI, it is quite handy.

On ChessGU 0.123, I've had some problems with the engine list, though: sometimes the list was completely erased, other, duplicates of all or almost all engines were created, e.g.

Alaric 707 :: Flags = UCI2 Can_Analyze Remis_Book
Alaric707 :: Flags = No_GUI_Book

the first is the one created by me, the second appeared after a restart of ChessGUI.

Also, I think Glaurung had problems, I had to install more than once, when trying to set up sometimes ChessGui would query and Glaurung not respond, and finally, the first tournament Glaurung lost all its games, but blundering, not being beaten normally.

I can try to provide more information, if you need.

anyway, thanks, nice gui.
horacio

Hi Horacio,
this all sounds like too many problems ;(
If they persist, consider a clean re-installation of ChessGUI or a manual fixing of ChsGUI_Eng.TXT .
Although I saw ChessGUI 0.124 exit unexpectedly, I still recommend it.
I am running it with more than 1100 working engines.
Feedback is always good.
Regards,
Matthias.

Matthias Gemuh wrote:ChessGUI 0.124 has just been "upped".

I have not tested it, but this version should be portable.
For portability :
- first install ChessGUI, engines and tablebases all in one hd partition.
- use uci engines which, apart from their own folders, only use the NalimovPath got from GUI.
- use wb engines which use only their own folders.
- portable engines must be marked in ChessGUI as portable.
- installed ChessGUI, engines and tablebases can then be copied to stick
- ChessGUI ships with a full set of 3-men tablebases
- on remote computer, tell ChessGUI in Preferences to work in portable mode and set NalimovPath only in GUI Preferences.

If it works for anybody, please report.

I have seen ChessGUI 0.123 and ChessGUI 0.124 exit unexpectedly in the interval between games. Maybe it affects only swiss tournaments ?
Watch out and report !

Matthias.

Oh la la !
ChessGUI 0.124 forgets to delete its debug files
You must delete them manually.
Bug will be fixed in next version.

Matthias.

Matthias Gemuh wrote:
I have seen ChessGUI 0.123 and ChessGUI 0.124 exit unexpectedly in the interval between games. Maybe it affects only swiss tournaments ?
Watch out and report !

Matthias.

No feedback yet

Please, report any unexpected ChessGUI exit in the interval between games.
Otherwise I will assume it occurs on my PC because of engine issues or is swiss-specific.

Thanx,
Matthias.

Matthias Gemuh wrote:

Matthias Gemuh wrote:
I have seen ChessGUI 0.123 and ChessGUI 0.124 exit unexpectedly in the interval between games. Maybe it affects only swiss tournaments ?
Watch out and report !

Matthias.

No feedback yet

Please, report any unexpected ChessGUI exit in the interval between games.
Otherwise I will assume it occurs on my PC because of engine issues or is swiss-specific.

Thanx,
Matthias.

Hi Matthias,

I played a 32-game match (gauntlet) between 2 engines (BugChess2 1.6.2 and Spike 1.2), no crash.

Gábor Szots wrote:
Hi Matthias,

I played a 32-game match (gauntlet) between 2 engines (BugChess2 1.6.2 and Spike 1.2), no crash.

Hi Gabor,
thanks for the feedback.
A bug that occurs only after hundreds of games is tough to locate.
However, there is hope because it occurs at a known point (between games).
I am already re-doing that interval.
Regards,
Matthias.

Hi Matthias,

No crash but I permanently find some engines not terminated after their game. They remain in memory and use a core of the CPU. This was not before 0.124.

Gábor Szots wrote:Hi Matthias,

No crash but I permanently find some engines not terminated after their game. They remain in memory and use a core of the CPU. This was not before 0.124.

I may decide to go back to ChessGUI 0.122 and continue from there .

Matthias.

Matthias Gemuh wrote:
Please, report any unexpected ChessGUI exit in the interval between games.
Otherwise I will assume it occurs on my PC because of engine issues or is swiss-specific.

Thanx,
Matthias.

ChessGUI 0.125

- known bugs fixed
- "Associate Icons" in menu

"Associate Icons" associates a bmp/jpg file with engine,
in the order :
- marching bmp
- marching jpg
- any bmp
- any jpg

Matthias.

Matthias Gemuh wrote:
ChessGUI 0.125

- known bugs fixed
- "Associate Icons" in menu

"Associate Icons" associates a bmp/jpg file with engine,
in the order :
- marching bmp
- marching jpg
- any bmp
- any jpg

Matthias.

Logos, not icons

Matthias Gemuh wrote:

Gábor Szots wrote:Hi Matthias,

No crash but I permanently find some engines not terminated after their game. They remain in memory and use a core of the CPU. This was not before 0.124.

I may decide to go back to ChessGUI 0.122 and continue from there .

Matthias.

Engines I find lurking in memory after their game is over: Djinn 0.925x, RomiChess P3k 64-bit.

I let a Djinn gauntlet run overnight and task manager showed 6 or 7 instances of Djinn when I woke up! I used version 0.125.

Gábor Szots wrote:
Engines I find lurking in memory after their game is over: Djinn 0.925x, RomiChess P3k 64-bit.

I let a Djinn gauntlet run overnight and task manager showed 6 or 7 instances of Djinn when I woke up! I used version 0.125.

This is strange because the latest approach gives the engines more time to unload. In my 1100-engine tournament, the engines don't hang.

I will stop adding new functionality to ChessGUI till the present problems are solved.


Matthias.

Matthias Gemuh wrote:

Gábor Szots wrote:
Engines I find lurking in memory after their game is over: Djinn 0.925x, RomiChess P3k 64-bit.

I let a Djinn gauntlet run overnight and task manager showed 6 or 7 instances of Djinn when I woke up! I used version 0.125.

This is strange because the latest approach gives the engines more time to unload. In my 1100-engine tournament, the engines don't hang.

I will stop adding new functionality to ChessGUI till the present problems are solved.


Matthias.

Fortunately I've experienced no problems at all with engines "hanging", but I haven't run any gauntlets under ChessGUI yet, only tournaments.

Graham Banks wrote:

Matthias Gemuh wrote:
This is strange because the latest approach gives the engines more time to unload. In my 1100-engine tournament, the engines don't hang.

I will stop adding new functionality to ChessGUI till the present problems are solved.


Matthias.

Fortunately I've experienced no problems at all with engines "hanging", but I haven't run any gauntlets under ChessGUI yet, only tournaments.

ChessGUI still exits occasionally.
I am beginning to think that someone must run a tournament with more than 1000 engines or a swiss tournament to encounter the problem.
Maybe my 512 MB RAM is too small memory for my giant tourney.

Matthias.

Gábor Szots wrote:Engines I find lurking in memory after their game is over: Djinn 0.925x, RomiChess P3k 64-bit.

I let a Djinn gauntlet run overnight and task manager showed 6 or 7 instances of Djinn when I woke up! I used version 0.125.

BTW, many non-native engines hang in most GUIs if they lose on time.
Were some of these games losses on time ?

Matthias Gemuh wrote:

Gábor Szots wrote:Engines I find lurking in memory after their game is over: Djinn 0.925x, RomiChess P3k 64-bit.

I let a Djinn gauntlet run overnight and task manager showed 6 or 7 instances of Djinn when I woke up! I used version 0.125.

BTW, many non-native engines hang in most GUIs if they lose on time.
Were some of these games losses on time ?

Yes, Djinn lost twice on time. And it won 5 times on time. And there were 2 games which ended like this:

"Djinn 0.925x draws against LambChop 10.99 on time "

Gábor Szots wrote:
Yes, Djinn lost twice on time. And it won 5 times on time. And there were 2 games which ended like this:

"Djinn 0.925x draws against LambChop 10.99 on time "

OK, then that explains at least some of the hangings.
My impression is that ChessGUI is stricter with time management than Arena.
To draw on time means the engine that still has time also has less than mating material.

Matthias.

Matthias Gemuh wrote:

Gábor Szots wrote:
Yes, Djinn lost twice on time. And it won 5 times on time. And there were 2 games which ended like this:

"Djinn 0.925x draws against LambChop 10.99 on time "

OK, then that explains at least some of the hangings.
My impression is that ChessGUI is stricter with time management than Arena.
To draw on time means the engine that still has time also has less than mating material.

Matthias.

Here are the drawn on time end positions:

8/8/3B4/5k2/3b4/8/6K1/8 b - -
8/5K2/8/8/5k2/4n3/8/5B2 b - -

In the second, white can mate black if black helps. Djinn was black.

Gábor Szots wrote:

Matthias Gemuh wrote:
OK, then that explains at least some of the hangings.
My impression is that ChessGUI is stricter with time management than Arena.
To draw on time means the engine that still has time also has less than mating material.

Matthias.

Here are the drawn on time end positions:

8/8/3B4/5k2/3b4/8/6K1/8 b - -
8/5K2/8/8/5k2/4n3/8/5B2 b - -

In the second, white can mate black if black helps. Djinn was black.

I was aware of this but chose to ignore the unlikely help.
Is FIDE always right about chess ?
I will implement an option to strictly comply with FIDE in next version.

Matthias.