Source code needed for simple experiment

[Michael Sherwin]

Is source code available for a material only searcher that is simple fast and hopefully smp? I want to do a very simple experiment for a search based on statistics and material only with absolutely no positional evaluation involved.

[Dann Corbit]

Is source code available for a material only searcher that is simple fast and hopefully smp? I want to do a very simple experiment for a search based on statistics and material only with absolutely no positional evaluation involved.

Suggestion:
Start with Viper, and simplify the eval to have wood only.

[Michael Sherwin]

Is source code available for a material only searcher that is simple fast and hopefully smp? I want to do a very simple experiment for a search based on statistics and material only with absolutely no positional evaluation involved.

Suggestion:
Start with Viper, and simplify the eval to have wood only.

Thanks Dann, Sounds good! I recently downloaded Jim Ablett's Compile of Viper and could not get it to work in Windows 10 under Arena. I then tried compiling Viper myself with the same result. What am I doing wrong?

[mar]

Is source code available for a material only searcher that is simple fast and hopefully smp? I want to do a very simple experiment for a search based on statistics and material only with absolutely no positional evaluation involved.

What is simple fast searcher?
You can try my engine: https://github.com/kmar/cheng4
in eval.cpp, just replace the contents of Eval::eval() with a call to fastEval()
and in psq.cpp, simply zero all piece-square tables. That should do it.

[Michael Sherwin]

Is source code available for a material only searcher that is simple fast and hopefully smp? I want to do a very simple experiment for a search based on statistics and material only with absolutely no positional evaluation involved.

What is simple fast searcher?
You can try my engine: https://github.com/kmar/cheng4
in eval.cpp, just replace the contents of Eval::eval() with a call to fastEval()
and in psq.cpp, simply zero all piece-square tables. That should do it.

Thanks Martin, Will it compile with MSVC 2017 or MSVC 2005?

[mar]

Thanks Martin, Will it compile with MSVC 2017 or MSVC 2005?

Not sure if with 2005 (it may), but it comes with VS2012 solution, so it should compile just fine.

[Michael Sherwin]

Is source code available for a material only searcher that is simple fast and hopefully smp? I want to do a very simple experiment for a search based on statistics and material only with absolutely no positional evaluation involved.

What is simple fast searcher?
You can try my engine: https://github.com/kmar/cheng4
in eval.cpp, just replace the contents of Eval::eval() with a call to fastEval()
and in psq.cpp, simply zero all piece-square tables. That should do it.

Thanks Martin, Will it compile with MSVC 2017 or MSVC 2005?

Martin, I'm old tired and not in very good health. Copying and pasting 61 files, getting them set up and hoping for no compatibility issues between MSVC 2012 and the versions of MSVC that I have sounds like a recipe for frustration. Isn't there a zip file of your source. I'm willing to do the work. However if there is a material searcher available the entire experiment would only take an hour at most to implement. Maybe you are willing to do the experiment? The very first experiment is so simple. Just count all the beta cuts favorable to the engine for each move at root and pick the best material move. If there are two or more materially equal moves pick the move with the highest beta cut count.

[mar]

Martin, I'm old tired and not in very good health. Copying and pasting 61 files, getting them set up and hoping for no compatibility issues between MSVC 2012 and the versions of MSVC that I have sounds like a recipe for frustration. Isn't there a zip file of your source. I'm willing to do the work. However if there is a material searcher available the entire experiment would only take an hour at most to implement. Maybe you are willing to do the experiment? The very first experiment is so simple. Just count all the beta cuts favorable to the engine for each move at root and pick the best material move. If there are two or more materially equal moves pick the move with the highest beta cut count.

It's simple, just click "clone or download" button and click download zip.
Then drag and drop the sln file to VS2017 and let it upgrade.
Release/x64 is probably the configuration you want.

[Michael Sherwin]

Martin, I'm old tired and not in very good health. Copying and pasting 61 files, getting them set up and hoping for no compatibility issues between MSVC 2012 and the versions of MSVC that I have sounds like a recipe for frustration. Isn't there a zip file of your source. I'm willing to do the work. However if there is a material searcher available the entire experiment would only take an hour at most to implement. Maybe you are willing to do the experiment? The very first experiment is so simple. Just count all the beta cuts favorable to the engine for each move at root and pick the best material move. If there are two or more materially equal moves pick the move with the highest beta cut count.

It's simple, just click "clone or download" button and click download zip.
Then drag and drop the sln file to VS2017 and let it upgrade.
Release/x64 is probably the configuration you want.

Okay thanks, I might be getting old but I can still learn new tricks!

[mar]

one more thing: it seems Microsoft broke the CRT (still not fixed!) so you'll probably have to set Windows SDK to 10.x in general properties or you can simply go to main.cpp and delete the 4 lines below "disable I/O buffering".

have fun