Winboard Forum

by **Tord Romstad** » 21 Aug 2004, 10:32

Geschrieben von:/Posted by: Tord Romstad at 21 August 2004 11:32:34:

Hi everybody,
You can now download Glaurung 0.1.2, a new UCI chess engine, from the link given below.
It is a very small and simple engine, comparable to TSCP or Gerbil (but it is of course not
a clone of either of these engines). At the moment, Glaurung is a rather weak engine,
but it will improve. I release it at this early stage because weak chess engines are usually
more fun than their stronger cousins.
Glaurung does not yet have any opening book code. The download does not contain a
readme file, but I think the engine will be easy to install in any UCI compatible GUI. It
also works well in xboard with the PolyGlot adapter.
Have fun!
Tord

Glaurung

by **Tord Romstad** » 21 Aug 2004, 10:33

Geschrieben von:/Posted by: Tord Romstad at 21 August 2004 11:33:31:
Als Antwort auf:/In reply to: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Tord Romstad at 21 August 2004 11:32:34:

I forgot:
Thanks a lot to Dann Corbit for compiling the Windows executable of Glaurung!
Tord

by **Patrick Buchmann** » 21 Aug 2004, 12:41

Geschrieben von:/Posted by: Patrick Buchmann at 21 August 2004 13:41:49:
Als Antwort auf:/In reply to: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Tord Romstad at 21 August 2004 11:32:34:

Hi everybody,
You can now download Glaurung 0.1.2, a new UCI chess engine, from the link given below.
It is a very small and simple engine, comparable to TSCP or Gerbil (but it is of course not
a clone of either of these engines). At the moment, Glaurung is a rather weak engine,
but it will improve. I release it at this early stage because weak chess engines are usually
more fun than their stronger cousins.
Glaurung does not yet have any opening book code. The download does not contain a
readme file, but I think the engine will be easy to install in any UCI compatible GUI. It
also works well in xboard with the PolyGlot adapter.
Have fun!
Tord

Hi Tord,
Thanks for the new UCI Engine.
Patrick

UCI Engines Ligue

by **Geoff Westwood** » 21 Aug 2004, 15:00

Geschrieben von:/Posted by: Geoff Westwood at 21 August 2004 16:00:08:
Als Antwort auf:/In reply to: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Tord Romstad at 21 August 2004 11:32:34:

Hello Tord
Thanks for the release of Glarung, it will be interesting to try it out and see how it does against my program.

It is a very small and simple engine, comparable to TSCP or Gerbil (but it is

The analogy to TSCP is a little tenuous. On a very quick test Glarung is searching tremendously deeply with a very small figure for nodes to depth !
I improved my program in that respect from TSCP, but Glaurung is outsearching me consistently by 2 or 3 ply in the opening and middle game.
Glaurung looks like a scarily efficiently starting point, it will be a killer in a year or so's time when you have added some more chess knowledge.
It is interesting to try it with the new version of Arena, (Seems to work fine in that by the way). This new version prints out the nodes searched at each depth in the main analysis window. This new feature certainly confirmed I need to get my nodes/to depth seriously reduced to be competetive, I was 20 fold worse in some positions :-(

At least my nodes/sec count is better than most so that compensates a little
Any insights into how you have got that figure so low would be interesting to read.
Anyways I will go and check see how the new Arena and Glarung are doing now, thanks again for the early release.
Regards Geoff

by **Uri Blass** » 21 Aug 2004, 16:31

Geschrieben von:/Posted by: Uri Blass at 21 August 2004 17:31:07:
Als Antwort auf:/In reply to: Re: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Geoff Westwood at 21 August 2004 16:00:08:

Hello Tord
Thanks for the release of Glarung, it will be interesting to try it out and see how it does against my program.
It is a very small and simple engine, comparable to TSCP or Gerbil (but it is
The analogy to TSCP is a little tenuous. On a very quick test Glarung is searching tremendously deeply with a very small figure for nodes to depth !
I improved my program in that respect from TSCP, but Glaurung is outsearching me consistently by 2 or 3 ply in the opening and middle game.
Glaurung looks like a scarily efficiently starting point, it will be a killer in a year or so's time when you have added some more chess knowledge.
It is interesting to try it with the new version of Arena, (Seems to work fine in that by the way). This new version prints out the nodes searched at each depth in the main analysis window. This new feature certainly confirmed I need to get my nodes/to depth seriously reduced to be competetive, I was 20 fold worse in some positions
At least my nodes/sec count is better than most so that compensates a little

Nodes/depth is not a good comparison because if you add pruning you improve nodes/depth and even if the pruning is productive nodes/depth can give wrong picture for how much it is productive.

Nodes/second is also bad comparison because by adding pruning rules you may reduce nodes/second count and increase the level of the program.

Uri

by **Tord Romstad** » 21 Aug 2004, 18:23

Geschrieben von:/Posted by: Tord Romstad at 21 August 2004 19:23:48:
Als Antwort auf:/In reply to: Re: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Geoff Westwood at 21 August 2004 16:00:08:

Hello Joff!

Hello Tord
Thanks for the release of Glarung, it will be interesting to try it out and see how it does against my program.
It is a very small and simple engine, comparable to TSCP or Gerbil (but it is
The analogy to TSCP is a little tenuous. On a very quick test Glarung is searching tremendously deeply with a very small figure for nodes to depth !
I improved my program in that respect from TSCP, but Glaurung is outsearching me consistently by 2 or 3 ply in the opening and middle game.
Glaurung looks like a scarily efficiently starting point, it will be a killer in a year or so's time when you have added some more chess knowledge.
It is interesting to try it with the new version of Arena, (Seems to work fine in that by the way). This new version prints out the nodes searched at each depth in the main analysis window. This new feature certainly confirmed I need to get my nodes/to depth seriously reduced to be competetive, I was 20 fold worse in some positions
At least my nodes/sec count is better than most so that compensates a little
Any insights into how you have got that figure so low would be interesting to read.
Anyways I will go and check see how the new Arena and Glarung are doing now, thanks again for the early release.

For the record: I didn't mean that Glaurung is comparable to TSCP or Gerbil
in strength, but only in the complexity of the code. I am almost willing to
bet that Glaurung is at least a couple of hundred rating points stronger
than TSCP. I am running a Noomen match between the two engines right now.
Glaurung won the first 11 games, and has a clear advantage in the 12th.
But does this mean simply that Glaurung displays a search depth 2 or 3
plies higher, or that it consistently seems to find the important moves
2 or 3 plies earlier? If it is the former, you have nothing to worry
about.
Well, I hope it will eventually be stronger than Gothmog, but there's a
very long way to go.
This is a good plan, but be careful about comparing with other engines.
All engines have different pruning and extension schemes, and n plies in
engine X usually isn't equivalent to n plies in engine Y. It is better
to compare to earlier versions of your own engine.
I don't have the time to give any detailed explanations right now, but
I'll try to do it later. Perhaps I will even make the source code public.
You're welcome! :-)

Tord

by **Tom Likens** » 21 Aug 2004, 18:34

Geschrieben von:/Posted by: Tom Likens at 21 August 2004 19:34:32:
Als Antwort auf:/In reply to: Re: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Tord Romstad at 21 August 2004 19:23:48:

Hello Tord,
I'm at work today and can't download your new engine. So what was the final
verdict, did you release the source or not?
regards,
--tom

by **Tord Romstad** » 21 Aug 2004, 18:37

Geschrieben von:/Posted by: Tord Romstad at 21 August 2004 19:37:45:
Als Antwort auf:/In reply to: Re: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Tom Likens at 21 August 2004 19:34:32:

Hello Tord,
I'm at work today and can't download your new engine. So what was the final
verdict, did you release the source or not?

I haven't decided yet. At the moment, the source is not available for
download, but it is still possible that I will release it later.
Tord

by **Dieter Eberle** » 21 Aug 2004, 20:08

Geschrieben von:/Posted by: Dieter Eberle at 21 August 2004 21:08:25:
Als Antwort auf:/In reply to: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Tord Romstad at 21 August 2004 11:32:34:

Hi everybody,
You can now download Glaurung 0.1.2, a new UCI chess engine, from the link given below.
It is a very small and simple engine, comparable to TSCP or Gerbil (but it is of course not
a clone of either of these engines). At the moment, Glaurung is a rather weak engine,
but it will improve. I release it at this early stage because weak chess engines are usually
more fun than their stronger cousins.
Glaurung does not yet have any opening book code. The download does not contain a
readme file, but I think the engine will be easy to install in any UCI compatible GUI. It
also works well in xboard with the PolyGlot adapter.
Have fun!
Tord

Hi Tord,
at level 40/5 40/5 rest 5 and level 40/5 Glaurung plays nonsense (lots of king moves and sacrifices) under Arena 1.09. Are these time controls not implemented?
Regards Dieter

by **David Dahlem** » 21 Aug 2004, 20:15

Geschrieben von:/Posted by: David Dahlem at 21 August 2004 21:15:54:
Als Antwort auf:/In reply to: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Tord Romstad at 21 August 2004 11:32:34:

Hi everybody,
You can now download Glaurung 0.1.2, a new UCI chess engine, from the link given below.
It is a very small and simple engine, comparable to TSCP or Gerbil (but it is of course not
a clone of either of these engines). At the moment, Glaurung is a rather weak engine,
but it will improve. I release it at this early stage because weak chess engines are usually
more fun than their stronger cousins.
Glaurung does not yet have any opening book code. The download does not contain a
readme file, but I think the engine will be easy to install in any UCI compatible GUI. It
also works well in xboard with the PolyGlot adapter.
Have fun!
Tord

Hi Tord
I have a problem with Glaurung in Arena. The engine plays instantly, only reaching a depth of 1. Here is a sample game, and the Arena debug for this game. What am i doing wrong?
[Event "Arena Chess"]
[Site "?"]
[Date "2004.08.21"]
[Round "?"]
[White "Glaurung"]
[Black "Gothmog"]
[Result "0-1"]
[Time "14:02:26"]
[TimeControl "40/300:40/300:40/300"]
[Termination "normal"]
1.e4 c6 2.d3 d5 {-0.50/10 11} 3.exd5 {-0.18/1} cxd5 {+0.12/11 11} 4.Ke2 {-0.45/1} Bg4+ {+0.78/11 11} 5.Ke3 {-14.55/1} Bxd1 {0.00/11 11} 6.Kf4 {-16.95/1} Qc7+ {+22.71/8 8} 7.Kg5 {-17.42/1} Qe5+ {+M2/3} 8.Kh4 {-17.04/1} Qh5+ {+M1/3} 9.Kg3 {-16.27/1} {White resigns} 0-1
4112**-------------NewGame!!!-------------
7855*1*-------------------------Starting engine 1 Glaurung-------------------------
7855*1*Configured Engine 1 Type: UCI
7995*1*Engine 1 dir: C:\Arena\Engines\Glaurung
8040*1*Engine 1 commandline: C:\Arena\Engines\Glaurung\Glaurung.exe
8375>1:uci
84952:hard
12336>2:easy
12336>2:ping 1
123572:computer
12437>2:ics -
12523*1*Start calc, move no: 0
12527*1*MainBook.abk
13268**New move before message, game, Game: 0 Move: 1
13268**New move after message, game 0, should be: 0 engine to move:2
13269>2:ping 2
132891:isready
139902:ping 3
145032:level 40 5 0
14797>2:new
14797>2:random
14797>2:ics -
14798>2:level 40 5 0
14858>2:post
14858>2:hard
14858>2:easy
14858>2:ping 4
148792:computer
15019>2:force
15219>2:usermove e2e4
15219>2:usermove c7c6
15219>2:usermove d2d3
152402:go
15485

by **Tord Romstad** » 21 Aug 2004, 20:32

Geschrieben von:/Posted by: Tord Romstad at 21 August 2004 21:32:14:
Als Antwort auf:/In reply to: Re: OT: Glaurung 0.1.2 released geschrieben von:/posted by: Dieter Eberle at 21 August 2004 21:08:25:

Hi Tord,
at level 40/5 40/5 rest 5 and level 40/5 Glaurung plays nonsense (lots of king moves and sacrifices) under Arena 1.09. Are these time controls not implemented?

Sorry, I forgot to mention this. X moves/Y minutes time controls are not yet supported
(I never use such time controls myself). This will be added in the next version.
Thanks for reminding me!
Tord

by **Tord Romstad** » 21 Aug 2004, 20:36

Geschrieben von:/Posted by: Tord Romstad at 21 August 2004 21:36:28:
Als Antwort auf:/In reply to: Re: Glaurung plays instantly geschrieben von:/posted by: David Dahlem at 21 August 2004 21:15:54:

Hi Tord
I have a problem with Glaurung in Arena. The engine plays instantly, only reaching a depth of 1. Here is a sample game, and the Arena debug for this game. What am i doing wrong?

Hi Dave,
You do nothing wrong. See my reply to Dieter. Glaurung does not yet support
X moves/Y minutes time controls. You have to use sudden death time controls (with or
without increments).
Tord

Winboard Forum

Who is online