Re: PolyGlot

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 30 March 2004 12:01:25:
Als Antwort auf:/In reply to: Re: PolyGlot geschrieben von:/posted by: Fabien Letouzey at 30 March 2004 10:21:57:

Cassandre, EnginMax 4.92(setoption name OwnBook value true), EnginMax 5.11c, FUSC#, Gaia, Hagrid, Patzer, Piranha, Silke Chess, T.rex, WJChess, Yace not works and for all the error message is:
"< ENGINE bestmove d2d4 file "xboard.cpp", line 689, assertion "BestMove!=MoveNone" failed"

Just to make sure the INI file is correct, the syntax in the [Engine] section should be: "OwnBook = true"; no "setoption", "name" or "value" in that file.
I tried to mimic the Windows INI file syntax although I am not very familiar with it. In particular I am not trying to make PolyGlot compatible with UCI2WB RC files, sorry.
Fabien.

[Günther Simon]

Geschrieben von:/Posted by: Günther Simon at 30 March 2004 16:05:41:
Als Antwort auf:/In reply to: Re: PolyGlot geschrieben von:/posted by: Fabien Letouzey at 30 March 2004 14:48:10:

Yesterday I have tested also Celes, Queen, Tao, Sjeng, Deep Sjeng, Big Lion, Taktix, Frenzee, Gandalf, Leila, Madeleine, and down there are the results. Note that Frenzee don't Knows a command "position startpos"
Works fine these:
Abrok, AnMon, Aristarch, Big Lion, Deep Sjeng, DelphiMax, Fruit, Gandalf, Hermann 1.0, Leila, Little Goliath, Madeleine, Nejmet, Pharaon, Queen, Sjeng, Taktix
Dragon, Shredder not works and for all the error message is:
"< ENGINE info currmovenumber 3 currmove d2d4 file "xboard.cpp", line 966, assertion "RootMove!=MoveNone" failed"

for Tao the error message is:
"< ENGINE info depth 6 seldepth 15 currmovenumber 5 currmove e2e3 file "xboard.cpp", line 966, assertion "RootMove!=MoveNone" failed"

Cassandre, Celes, EnginMax 4.92(setoption name OwnBook value true), EnginMax 5.11c, FUSC#, Gaia, Hagrid, Patzer, Piranha, Silke Chess, T.rex, WJChess, Yace not works and for all the error message is:
"< ENGINE bestmove d2d4 file "xboard.cpp", line 689, assertion "BestMove!=MoveNone" failed"

For EnginMax 4.92(setoption name OwnBook value false) the error message is:
"< ENGINE info depth 1 seldepth 0 score cp 42 time 29 nodes 25 pv d2d4 file "xboard.cpp", line 908, assertion "seldepth>0" failed"

For SOS 4 for Arena the error message is:
"< ENGINE info depth 1 seldepth 0 file "xboard.cpp", line 908, assertion "seldepth>0" failed"

For Hermann 1.0(setoption name OwnBook value true) the error message is:
"< ENGINE info depth 0 seldepth 1 time 20 nodes 0 hashfull 0 nps 0 score cp 1 file "xboard.cpp", line 807, assertion "depth>0" failed
For Frenzee 151 the error message is:
"< ENGINE Error (unknown command): position startpos < XBOARD ?"
Best Regards
Salvo Spitaleri

It's starting to look better
Hm ... I think, except for Hermann, the updated version should fix those problems. Or are you reporting problems with the latest version (see Dann's post date).
It would be interesting to have a look at Frenzee's log file.
But I am making more changes to PolyGlot that are related to I/O so it is better to wait for an update by Dann in a few hours, and see which problems still occur.
Thanks a lot!!!
Fabien.

Just a note: IMHO its not necessary to test those, which already
support UCI _and_ Winboard protocols.
In my last test yesterday night all programs I tested worked now!
The last problems remaining seem to be proper unloading, like
Dann mentioned and the CPU load.
At least on my test machine Polyglot used between 5-10% of CPU time,
when the UCI engine was to move.
It's possible though that this is a Win98 issue, as Igor described
the same syptom with the same OS.
In the next couple of days I am going to test on my main machine
with WINXP together with Joris WBTM, which is very sensible for
not properly unloading of programs.
Best regards,
Günther

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 30 March 2004 16:13:22:
Als Antwort auf:/In reply to: Re: PolyGlot geschrieben von:/posted by: G�nther Simon at 30 March 2004 16:05:41:

Just a note: IMHO its not necessary to test those, which already
support UCI _and_ Winboard protocols.
In my last test yesterday night all programs I tested worked now!
The last problems remaining seem to be proper unloading, like
Dann mentioned and the CPU load.
At least on my test machine Polyglot used between 5-10% of CPU time,
when the UCI engine was to move.
It's possible though that this is a Win98 issue, as Igor described
the same syptom with the same OS.
In the next couple of days I am going to test on my main machine
with WINXP together with Joris WBTM, which is very sensible for
not properly unloading of programs.

For you maybe, but I am interested in testing all UCI engines, not just UCI-only engines. Note that this is not as urgent maybe. But when most of UCI-only engines work fine, I will be interested in more tests, always.
Good!
They are unrelated issues. This week is about compatibility. PolyGlot has unoptimised I/O and is also running in DEBUG mode. We will see about speed when engines work fine.
I am interested in the "unload" problem of course, but I need log files to know what the problem actually is. For that PolyGlot log and also engine log is best.
OK.
Fabien.