C++ UCI Chess engine inspired by Stockfish. Created using ideas learned during the development of Boxfish.

Example of the engine running here. (built to WebAssembly with Emscripten, note there is ~10x speed decrease compared to native code)

Storm currently can use the same NNUE as Stockfish (future plans to train network independently). For now you should download a network from https://tests.stockfishchess.org/nns and copy it into the working directory before you run the program. Storm will default to a classical evaluation if no network file is found.

• Bitboards and magic bitboard move generation
• UCI protocol
• Search:
  • Transposition table with Zobrist hashing
  • PVS search
  • Aspiration windows and iterative deepening
  • Razoring
  • Adaptive null move pruning
  • Futility pruning
  • Late move reduction
• Evaluation:
  • NNUE
• All perft tests passed
• Pondering
• SEE move ordering

1. Download or clone this repository (use flag --recurse-submodules or --recursive to include submodules).
2. If on windows run the Scripts/Win-GenProjects.bat script to generate the Visual Studio 2019 project and solution files.
3. For best performance, download a compatible NNUE from here and copy into working directory. Working network nn-76a8a7ffb820.nnue. Storm will default to a classical evaluation if no valid network is found.

1. Run Scripts/Win-GenProjects.bat and build the solution using Visual Studio 2019.
2. Build outputs are located in the bin directory.

1. Run Scripts/Linux-GenProjects.sh to generate the Makefiles.
2. Run make -j<number_of_cores> Storm-Cli config=dist to build Storm.
3. Build outputs are located in the bin directory.

On linux, must use the config=distshared when building.

1. Install SWIG
2. Setup a SwigConfigWindows.lua or SwigConfigLinux.lua depending on your platform. (Copy from SwigConfig.example.lua)
3. Build the Storm-Swig project as above for your platform

1. Install emsdk
2. If using NNUE, copy your NNUE network into the Storm-Emscripten directory
3. Run emsdk activate then inside Storm-Emscripten run python build.py
4. Outputs are located in bin/Web