A chess engine written in C++, powered by magic bitboard and neural networks.

While this project is still a major WIP, the end-goal is a dual-mode engine:

• An iterative search engine with Alpha-Beta Pruning, Quiescence, etc.
• A neural network-powered engine using Monte Carlo Tree Search (MCTS)

For a quick build of the project, run:

git clone https://github.com/Water-Engine/water.git
cd water
make -j4 run

The engine communicates through the UCI protocol for terminal interaction.

• g++ with C++20
• GNU Make
• Clang-format
• Catch2 (included in this repository)
• cloc (for cloc make target > optional)
• python for script running

The project's build system uses C++20 and GNU Make, and it is recommended that you run make with the flag -j4 to run batch jobs. Below is a list of targets with their requirements where applicable:

• default: Builds the release configuration (default)
• install: Builds the dist config (to be updated)
• all: Builds all optimization configurations for the project (dist, release, and debug)
• dist: Builds the project with maximum optimization and disabled profiling
• release: Builds the project with slightly fewer optimizations and no DEBUG define
• debug: Builds the project with no optimization, defining both PROFILE and DEBUG
• test: Run the project's unit tests Excludes perft testing
• perft: Run the perft tests
• run: Alias for run-release
• run-dist: Build and run the dist binary
• run-release: Build and run the release binary
• run-debug: Build and run the debug binary
• fmt: Format all C++ source and header files using clang-format
• fmt-check: Validates C++ formatting rules without altering project files
• clean: Remove object files, dependency files, and binaries

• cloc: Count the lines of code in the project's relevant directories
• help: Print this help menu

Contributing guidelines, information on tests, formatting, and profiling can be found in CONTRIBUTING.md. You can also check out a WIP roadmap for the project at ROADMAP.md.

• Learn engine internals such as move generation, evaluation, and search
• Work on low-level performance optimizations in modern C++
• Explore and improve machine learning fine-tuning for chess