Glaurung 2.2

[Tord Romstad]

Hi all,

The source code and Mac OS X executables for Glaurung 2.2 are now available from my web site. Executables for GNU/Linux and Windows will follow shortly.

This version is mostly the result of work I did in preparation for the Open Polish Computer Chess Championships in June. When I tested the most recent development version before the tournament, it performed poorly, and I ended up using version 2.1 in the tournament. Evidently, I didn't test thoroughly enough, because recent tests indicate that the version I prepared for the OPCCC is considerably stronger than 2.1. I have therefore fixed a handful of very minor bugs, polished the code a little, and published the result as Glaurung 2.2.

The most important improvement is the addition of space evaluation. This improves the strength (particularly in Chess960, but also in normal chess), and also makes the style of play more attractive.

For the next version, I will have a close look at Marco Costalba's improvements in Stockfish 1.1. Stockfish 1.1 seems extremely strong (possibly stronger than Glaurung 2.2), and it is surprising that it doesn't get more attention. In the current version of Glaurung, there is only one new thing stolen from Stockfish: Razoring has been modified to use only the approximate evaluation of the position, rather than the full eval.

Tord

[Volker Pittlik]

Hi all,

The source code and Mac OS X executables for Glaurung 2.2 are now available from my web site...

Very nice Christmas present, Tord.

Thank you very much!


Volker

[Tord Romstad]

The source code and Mac OS X executables for Glaurung 2.2 are now available from my web site. Executables for GNU/Linux and Windows will follow shortly.

The Windows binaries are now ready. The Linux binaries will follow later today.

Tord

[F.Huber]

The Windows binaries are now ready.

Thanks for the new Glaurung version, Tord!

Unfortunately an old problem of previous versions has come back again: the Windows executable is running slow like a snail on Windows 98 (about 15-20 times slower than version 2.1).

Exactly this problem occured already with a previous version of Glaurung (I believe it was 2.0), but it has been solved with 2.1 - and now it´s back again with 2.2.

IIRC it had something to do with the different threads Glaurung creates at starttime (or just any non-optimal compiler setting).
Maybe you remember how you´ve fixed this problem for the previous version 2.1?

Regards,
Franz

[Tord Romstad]

The Windows binaries are now ready.

Thanks for the new Glaurung version, Tord!

Unfortunately an old problem of previous versions has come back again: the Windows executable is running slow like a snail on Windows 98 (about 15-20 times slower than version 2.1).

Exactly this problem occured already with a previous version of Glaurung (I believe it was 2.0), but it has been solved with 2.1 - and now it´s back again with 2.2.

IIRC it had something to do with the different threads Glaurung creates at starttime (or just any non-optimal compiler setting).
Maybe you remember how you´ve fixed this problem for the previous version 2.1?

I'm sorry, but I have absolutely no idea. It probably wasn't me who fixed the problem back then: I don't do the Windows compiles, and I know close to nothing about Windows. I suppose it's a question of using the right compiler and/or compiler settings, but I can't give any more specific advice than this.

For Linux users, I have now added 32-bit and 64-bit binaries compiled in Ubuntu 8.10. I am not sure whether they work in other Linux distributions. I have also added a few missing #include directives in the source code, which turned out to be necessary in order to make the program compile with recent versions of GCC under Linux.

Tord

[Jim Ablett]

The Windows binaries are now ready.

Thanks for the new Glaurung version, Tord!

Unfortunately an old problem of previous versions has come back again: the Windows executable is running slow like a snail on Windows 98 (about 15-20 times slower than version 2.1).

Exactly this problem occured already with a previous version of Glaurung (I believe it was 2.0), but it has been solved with 2.1 - and now it´s back again with 2.2.

IIRC it had something to do with the different threads Glaurung creates at starttime (or just any non-optimal compiler setting).
Maybe you remember how you´ve fixed this problem for the previous version 2.1?

Regards,
Franz

Hi Franz,

It's a compiler problem. I switched to using MSVC compiler for the Glaurung 32 bit windows builds. They are a little faster than using
Intel compiler on my 32 bit Athlon running XP pro.

Here' an Intel compiler 32 bit build. This has been tested in a Win98 system and is working correctly.

http://www.mediafire.com/?cy1cw12jemh

Jim.

[F.Huber]

Hi Franz,

It's a compiler problem. I switched to using MSVC compiler for the Glaurung 32 bit windows builds. They are a little faster than using
Intel compiler on my 32 bit Athlon running XP pro.

Here' an Intel compiler 32 bit build. This has been tested in a Win98 system and is working correctly.

http://www.mediafire.com/?cy1cw12jemh

Jim.

Many thanks, Jim!

It´s really great that someone like you is still taking care of computer oldies like me , also your lastest Crafty builds were the only ones which are also running on Win98!

Thanks again for your great work,
Franz.

[nepossiver]

For the next version, I will have a close look at Marco Costalba's improvements in Stockfish 1.1. Stockfish 1.1 seems extremely strong (possibly stronger than Glaurung 2.2), and it is surprising that it doesn't get more attention.

CCRL 40/40 list has Stockfish 1.01 40-50 ELO weaker than Glaurung 2.1, and about same as Glaurung 2.01. Oddly enough, my testing show Stockfish 1.01 60 ELO stronger than Glaurung 2.1, and 100 over both Stockfish 1.0 & 1.1 - my games are short, though, mostly 1min+1s, or 40/4 repeating.

Anyway, I don't think strength is the reason Stockfish doesn't get much attention, it seems that "clones" are generally disliked, even legal forks of GPLed engines.

Anyway, thanks for Glaurung, it is a most entertaining engine to watch - it is quite an optimistic, in general !

[ChessKnight]

A quick questions for Glaurung 2.2

1) Which end-game database, Nalimov or Scorpio bit-bases, is compatible with Glaurung 2.2 ?

2) Did any body face a problem of engine shutdown (freezes) while using WinBoard + Polyglot + WinZippy + Glaurung 2.2 ? coz I found Glaurung2.2 freezes in somewhere in end-game and running out-of-time.

[Teemu Pudas]

1) Only its own KPK bitbase generated at startup.

[Tord Romstad]

A quick questions for Glaurung 2.2

1) Which end-game database, Nalimov or Scorpio bit-bases, is compatible with Glaurung 2.2 ?

Glaurung does not, and never will support Nalimov or Scorpio bitbases. The reason is that I did not write the Nalimov or Scorpio bitbase code, and I don't want to use somebody else's code in my own program.

As Teemu wrote, the only bitbase Glaurung currently uses is a KP vs K which is generated every time the program starts up. It is possible that I will add other bitbases later, but it doesn't have high priority at the moment. Glaurung plays a quite strong endgame without bitbases, its main weaknesses are in the opening and the middle game.

2) Did any body face a problem of engine shutdown (freezes) while using WinBoard + Polyglot + WinZippy + Glaurung 2.2 ? coz I found Glaurung2.2 freezes in somewhere in end-game and running out-of-time.

I've never seen it freeze or crash on my Mac, but I know nothing about Windows.

Tord

[ChessKnight]

As I asked you earlier

Quote:
2) Did any body face a problem of engine shutdown (freezes) while using WinBoard + Polyglot + WinZippy + Glaurung 2.2 ? coz I found Glaurung2.2 freezes in somewhere in end-game and running out-of-time.

It's not related to Glaurung, I guess its related to either polyglot or WinBoard coz not only Glaurung but same problem persist in other UCI engines too while using WinBoard + Polyglot + WinZippy + ICS. While in other GUIs all engines work fine.

But someone else in this forum also had a same problem of engine running out of time or freezes (sleep mode - without sending best move), and relative flag for the engine fells down and weaker engine wins the match.

Thanks n regards,