GULL 1.0 live on IPON

[IWB]

Hello

Normaly I am not testing anonymous engines, but the author is giving a name (Vadim Demichev) in the readme and it seems that he is considered as a real person by the community, so I thought I'll give it a try.

At the moment this is a test. If Gull 1.0 is stable in pondering I will put it into my list.

http://www.inwoba.de

Bye
Ingo

[Frank Quisinsky]

Hello Ingo,

I played yesterday 20 games with (game in 5 minutes) and 5 games (40 moves in 10 minutes repeadetly) on one of my Notebooks with ponder = on. I watched the time management, ponder and so on.

It seems that GullChess 1.0 works fine with ponder. No problems here. Time management is good too. I think the engine have around 2.725 - 2.800 ELO.

Thanks for your test ...
Let me look on your site now

Best
Frank

[IWB]

Hello Frank,

...
It seems that GullChess 1.0 works fine with ponder. No problems here. ...

Here too, more than 700 games and no crash!

... Time management is good too. ...

I dont mind the time mnagement. The engine is released and I test the playing strength. If the time management is bad it is weakening the engine which will be reflected in the rating - as losses on time are weakening an engine and should not be judeged by the "tester"!

I think the engine have around 2.725 - 2.800 ELO. ...

Quite a wide range. At the moment it seems to be somewhere between Hiarcs 13 and Deep Fritz 12, but I underestimated it as well and took opponents which are all "weaker" than gull. When the 1100 games match is finished I will play Gull vs the upper part of my list!

Bye
Ingo

[IWB]

Hello

Gull has finished its first 1900 games. Nice new entry at position no 9.

http://www.inwoba.de

Bye
Ingo

PS: A little change on the top has happend too

[gerold]

Hello

Gull has finished its first 1900 games. Nice new entry at position no 9.

http://www.inwoba.de

Bye
Ingo

PS: A little change on the top has happend too

Very good. Thanks for your testing. Much appreciated.

Best to you,
Gerold.

[IanO]

Gull is a relatively straightforward engine. I developed a patch to compile version 1.0a using gcc on x86_64 Mac OS X (and probably Linux too). With bsf/bsr but without prefetch, I'm getting about 2M nps (about 15 ply in 5 seconds in the middle game) on a 2.66 GHz Core 2 Duo. Bookless, preliminary strength is on par with Glaurung 2.2. Does that sound about the same as the Win64 builds?

Like Robbolito, Gull often blitzes out wins from drawn endings in extreme time pressure.