I have just released Marvin 6.1.0. Based on my testing it's about 80 elo stronger than version 6.0.0.
Changes:
* Use hybrid eval instead of pure NNUE
* Improvements to PVS
* Improvements to LMR
* New net based on Marvin 6.0.0 data
Thanks to everyone that keeps on testing my engine.
Binaries: https://github.com/bmdanielsson/marvin- ... tag/v6.1.0
Marvin 6.1.0
Moderators: hgm, Dann Corbit, Harvey Williamson
-
Martin
- Posts: 51
- Joined: Sun Jul 10, 2016 9:12 pm
- Location: Lund, Sweden
- Full name: Martin Danielsson
-
matejst
- Posts: 363
- Joined: Mon May 14, 2007 8:20 pm
- Full name: Boban Stanojević
Re: Marvin 6.1.0
Thanks, Martin. Will try it asap.
-
Graham Banks
- Posts: 41178
- Joined: Sun Feb 26, 2006 10:52 am
- Location: Auckland, NZ
-
Damir
- Posts: 2798
- Joined: Mon Feb 11, 2008 3:53 pm
- Location: Denmark
- Full name: Damir Desevac
Re: Marvin 6.1.0
Thanks Martin. Very nice progres. I hope you continue to improve your engine. 
-
chessica
- Posts: 572
- Joined: Thu Aug 11, 2022 11:30 pm
- Full name: Esmeralda Pinto
Re: Marvin 6.1.0
Hello Damir,
i wanted to compile marvin from sources and i get this error "pthread.h": No such file or directory from microsoft visual studio 2022.
Are the sources incomplete?
i wanted to compile marvin from sources and i get this error "pthread.h": No such file or directory from microsoft visual studio 2022.
Are the sources incomplete?
-
Martin
- Posts: 51
- Joined: Sun Jul 10, 2016 9:12 pm
- Location: Lund, Sweden
- Full name: Martin Danielsson
Re: Marvin 6.1.0
I use gcc and the included Makefile when building for Windows. Using Visual Studio should be possible, but it might take some work (I haven't tried it for a long time). To avoid the 'pthread.h' error you have to define the WINDOWS preprocessor variable. Additionally, I also think you will have to define _CRT_SECURE_NO_DEPRECATE in order to avoid some other issues.