Glaurung 2.2 - no EGTB usage?

[royb]

This is primarily intended for Tord but if others know the answer, so much the better ...

Does Glaurung 2.2 use EGTBs? I see no UCI option for it (no NalimovPath variable for example). There is a "SearchLog" mentioned but I do not know that is for or how it may be used. Any info on that is also appreciated.

Thanks,

Roy

[geots]

This is primarily intended for Tord but if others know the answer, so much the better ...

Does Glaurung 2.2 use EGTBs? I see no UCI option for it (no NalimovPath variable for example). There is a "SearchLog" mentioned but I do not know that is for or how it may be used. Any info on that is also appreciated.

Thanks,

Roy

No, Glaurung 2.2 does NOT use egtb.

[Tord Romstad]

This is primarily intended for Tord but if others know the answer, so much the better ...

Does Glaurung 2.2 use EGTBs? I see no UCI option for it (no NalimovPath variable for example).

Glaurung does not, and never will support Nalimov tablebases. For at least two reasons, I can't use them: The first reason is that I didn't write the code to generate and probe the Nalimov tablebases, and using somebody else's code would make my program a semi-clone. The second reason is that the license of Nalimov's EGTB code is incompatible with Glaurung's GNU General Public License, and I have no desire to drop the GPL.

There is a "SearchLog" mentioned but I do not know that is for or how it may be used. Any info on that is also appreciated.

That's simply a log file where Glaurung writes its main lines and evaluations. If you want to see how it looks, just switch on "Use Search Log", play a game or two, and inspect the file named "SearchLog.txt" in Glaurung's home directory. I'm not sure it's very useful for anyone except myself.

Tord

[govert]

Correct me if i'm wrong, but does not Crafty support nalimov, as well as being released under GPL?

[jwes]

Correct me if i'm wrong, but does not Crafty support nalimov, as well as being released under GPL?

Crafty supports nalimov, but is not released under GPL (it has it's own open source license).

[glorfindel]

Crafty's licence is not GPL, though its source code is available for everyone to study and modify for personal use. You can read the licence in the comments at the beginning of the file main.c of its source code.

[Dirt]

Glaurung does not, and never will support Nalimov tablebases. For at least two reasons, I can't use them: ... The second reason is that the license of Nalimov's EGTB code is incompatible with Glaurung's GNU General Public License, and I have no desire to drop the GPL.

In theory, you shouldn't need the Nalimov code to access the tablebases. If someone will document the tablebase format, then clean GPL code could be written to access them.

Now, I'm pretty sure you're not interested in such a project, and I haven't seen any evidence that anyone else is, either. In the off chance that such code is nonetheless written, then while Glaurung might not use it there would be nothing stopping Stockfish, or similar derivatives, from using it.

I find the Nalimov tablebases, which use DTM, to not be very interesting anyhow.

[bob]

Correct me if i'm wrong, but does not Crafty support nalimov, as well as being released under GPL?

No, crafty is not under "GPL".

It is open source, but not GPL open source.

[Tord Romstad]

Correct me if i'm wrong, but does not Crafty support nalimov, as well as being released under GPL?

As several others have pointed out, Crafty is not and never was released under the GPL, but it is not surprising that you thought it were: There is a very widespread myth that Crafty is a GPLed program, and you'll see it described as such many places. I wonder where this myth comes from? Could it be because it used to be the "standard chess program" in Linux, and/or because people confuse Crafty with GNUChess?

Tord

[Tord Romstad]

In theory, you shouldn't need the Nalimov code to access the tablebases. If someone will document the tablebase format, then clean GPL code could be written to access them.

Now, I'm pretty sure you're not interested in such a project, and I haven't seen any evidence that anyone else is, either. In the off chance that such code is nonetheless written, then while Glaurung might not use it there would be nothing stopping Stockfish, or similar derivatives, from using it.

You are right, writing original code for probing the Nalimov tablebases would be possible in principle, although I don't know how hard it would be. I suspect that writing completely new code for generating, indexing, compressing and probing tablebases wouldn't be that much harder, and it would certainly be a lot more interesting.

I find the Nalimov tablebases, which use DTM, to not be very interesting anyhow.

Personally I find tablebases in general to be of limited interest, at least with the current level of technology. I don't want to waste gigabytes of disk space for something that is worth at most a handful of Elo points.

Tord