Few errors around Igel engine used on CCRL website

[voffka]

Hello everyone,

I noticed that Igel engine has been used in CCRL tests as seen at http://computerchess.org.uk/ccrl/4040/c ... librate=no

There is one problem though, this version is declared at http://computerchess.org.uk/ccrl/4040/c ... 0#igel_1_0 as being developed by "Vladimir Medvedev", but in reality Igel is derived from Greko and developed by "Volodymyr Shcherbyna".

Another problem is that Igel 1.0 was never an official release, the first official release of engine is Igel 0.8 and it is available on it's official github page https://github.com/vshcherbyna/igel/releases

Can I please ask to update the details of author and project page please? also, would it be possible to change the name of Igel 1.0 used on the web-site to Igel 0.4 - I think it was the dev version used.

Last but not least, please use Igel 0.8 for further testing, it should be a stronger version as it adds lazy smp and other improvements.

Thanks!
Volodymyr

[Ray]

Hi,

Thanks for your posting.

Because Igel is a GreKo derivative and part of the GreKo family, Vladimir's name will always appear and there is nothing we can do about that. It is just how the website functions. But we will certainly add your name as well as the programmer of Igel. Apologies, it should have been there already but we made a mistake in the config files.

[voffka]

Hello Ray,

Thank you for the update. It sounds like a good plan

Regarding the versioning, I wonder if there is any chance that 1.0 would be changed to 0.4 on CCRL website? For further releases I would make sure that I will not use version 1.0 to create any further confusion.

So the 0.8 is the first official, later on the next would be either 0.9 or 1.1 to make it simple.

[Graham Banks]

Hello Ray,

Thank you for the update. It sounds like a good plan

Regarding the versioning, I wonder if there is any chance that 1.0 would be changed to 0.4 on CCRL website? For further releases I would make sure that I will not use version 1.0 to create any further confusion.

So the 0.8 is the first official, later on the next would be either 0.9 or 1.1 to make it simple.

I've asked Ray to rename 1.0 to 0.4.
Will test 0.8 soon.
Is there a 64-bit compile?

Graham.

[T-Poppins]

Hi voffka!

Thank you for bringing up this issue here. A couple of points need to be clarified in addition to those you brought up:

1) should the name of the engine be listed as "Igel" (capitalized) rather than the current "igel"?
2) what country should be listed along with your name?
3) if the 0.4 dev version was not distributed by you and you don't want it listed you can have it removed

Tirsa

[voffka]

Hello Graham,

I've asked Ray to rename 1.0 to 0.4.
Will test 0.8 soon.
Is there a 64-bit compile?

Graham.

Thanks!

The zip file at https://github.com/vshcherbyna/igel/releases contains:

Windows: x32 bit and x64
Linux: x32 bit and x64
Mac: x64

The 0.8 supports Threads option, therefore please use it if possible

[voffka]

Hello T-Poppins,

Hi voffka!

Thank you for bringing up this issue here. A couple of points need to be clarified in addition to those you brought up:

1) should the name of the engine be listed as "Igel" (capitalized) rather than the current "igel"?
2) what country should be listed along with your name?
3) if the 0.4 dev version was not distributed by you and you don't want it listed you can have it removed

Tirsa

1. Igel is fine
2. I am originally from Ukraine
3. yes, it was not distributed by me, so if possible please remove it. Instead it would be nice just to start with 0.8 official release from https://github.com/vshcherbyna/igel/releases

Thanks a lot!

[Graham Banks]

SendFirstCommands UCI
SendToEng2Time 0000001746865266 : Eng02 (Igel 0.8 64-bit) -> uci
.
Scheduling StartEngineTimerEvent() for Igel 0.8 64-bit at 0000001746865313.
Waiting for Igel 0.8 64-bit.
Eng02 (Igel 0.8 64-bit) <- igel 0.8 by V. Medvedev, V. Shcherbyna
Eng02 (Igel 0.8 64-bit) <- id name igel 0.8
nChessVariant(1) = 1, id name.
Cancelling StartEngineTimerEvent() for Igel 0.8 64-bit at 0000001746865469.
Eng02 (Igel 0.8 64-bit) <- id author V. Medvedev, V. Shcherbyna
Eng02 (Igel 0.8 64-bit) <- option name Hash type spin default 128 min 1 max 16384
Eng02 (Igel 0.8 64-bit) <- option name Threads type spin default 1 min 1 max 128
Eng02 (Igel 0.8 64-bit) <- option name Strength type spin default 100 min 0 max 100
Eng02 (Igel 0.8 64-bit) <- uciok
nChessVariant(1) = 1, uciok.
SendToEng2Time 0000001746865938 : Eng02 (Igel 0.8 64-bit) -> setoption name Hash value 256
.
SendToEng2Time 0000001746866063 : Eng02 (Igel 0.8 64-bit) -> setoption name Threads value 1
.
SendToEng2Time 0000001746866141 : Eng02 (Igel 0.8 64-bit) -> setoption name Strength value 100
.
Total Physical Memory = 4095 MB.
Total Available Memory = 4095 MB.
Hash + TbCache to allocate for Igel 0.8 64-bit = 288 MB.
SendToEng2Time 0000001746866235 : Eng02 (Igel 0.8 64-bit) -> isready
.
Scheduling StartEngineTimerEvent() for Igel 0.8 64-bit at 0000001746866329.
Waiting for Igel 0.8 64-bit.
Eng02 (Igel 0.8 64-bit) <- Unknown option Strength
StartEngineTimeout of Igel 0.8 64-bit (Event Nr. 1) at Time0000001746871360.
Eng02 (Igel 0.8 64-bit) <- StartEngineTimedOut 1
nChessVariant(1) = 1, readyok.
SendToEngine(Igel 0.8 64-bit, ucinewgame
) failed : Process is dead !
SendToEngine(Igel 0.8 64-bit, isready
) failed : Process is undefined !

[voffka]

Hello Graham,

SendFirstCommands UCI
Eng02 (Igel 0.8 64-bit) <- Unknown option Strength

I removed option "Strength". Can you please try without it?

P.S. In either case it should not crash if an unknown option is passed. Can you please tell me what tool you are using? I am mostly testing with cutechess-cli.

[Graham Banks]

Hello Graham,

SendFirstCommands UCI
Eng02 (Igel 0.8 64-bit) <- Unknown option Strength

I removed option "Strength". Can you please try without it?

P.S. In either case it should not crash if an unknown option is passed. Can you please tell me what tool you are using? I am mostly testing with cutechess-cli.

I use ChessGUI for all of my testing.
Will try it now.

[voffka]

Hello Graham,

I use ChessGUI for all of my testing.
Will try it now.

I've never tried this tool, but I will check it out and will update on what could be the problem.

[Graham Banks]

Tony Mokonen posted this in Talkchess:

The code for the Strength option has been removed from this GreKo derivative in search.cpp, and in the OnSetoption routine in main.cpp, but not from the output produced when the uci command is issued. Hence, the program exits when the GUI tries to set a value for Strength.

[voffka]

Hello Graham,

Tony Mokonen posted this in Talkchess:
The code for the Strength option has been removed from this GreKo derivative in search.cpp, and in the OnSetoption routine in main.cpp, but not from the output produced when the uci command is issued. Hence, the program exits when the GUI tries to set a value for Strength.

I don't think the problem is any latest changes I've done in Igel. I've tried ChessGUI with GreKo 2018.06 and the same problem happens there - it does not complain on Strengh, but the process still dies.

I will update when I find the culprit.

Update: my above observations were wrong. The root-cause was indeed an issue with Strength.
P.S. My other statements regarding GreKo are incorrect here as well - it works well with ChessGUI.

[voffka]

Hello Graham,

Can you please try Igel 1.1 from https://github.com/vshcherbyna/igel/releases and let me know if the problem is fixed?

Thanks in advance!

[Graham Banks]

Hello Graham,

Can you please try Igel 1.1 from https://github.com/vshcherbyna/igel/releases and let me know if the problem is fixed?

Thanks in advance!

Will try now.

[Graham Banks]

Works now. Gauntlet running in 16065.

[voffka]

Excellent. Thanks for the test!