Skip to content

Latest commit

 

History

History
141 lines (102 loc) · 4.52 KB

CONTRIBUTING.md

File metadata and controls

141 lines (102 loc) · 4.52 KB

Contributing to LeetGPU

Thank you for contributing to LeetGPU! We welcome new challenges and framework support.

What You Can Contribute

1. New Challenges

Add new GPU programming problems with starter templates across multiple frameworks.

2. New Framework Support

Add starter templates for new GPU programming frameworks to existing challenges.

3. Bug Fixes

Fix issues with existing starter templates or problem descriptions.

Adding a New Challenge

Each challenge should include:

challenges/<difficulty>/<number>_<name>/
├── challenge.html          # Problem description
├── challenge.py           # Reference implementation and metadata
└── starter/              # Starter templates
    ├── starter.cu           # CUDA 
    ├── starter.mojo         # Mojo 
    ├── starter.pytorch.py   # PyTorch 
    ├── starter.tinygrad.py  # TinyGrad 
    └── starter.triton.py    # Triton

Requirements

  • Clear problem description with 1 or more examples
  • Starter templates should follow the format of existing starter templates
  • Follow existing challenge patterns

Starter Template Guidelines

  • Comments: Match the format of existing challenges
  • Easy problems: Provide more starter code and helpful structure
  • Medium/Hard problems: Just have an empty solve function
  • All templates should compile but not solve the problem

Code Style and Linting

All starter code must pass automated linting checks before being merged. We use:

Automated Pre-commit Hooks (Recommended)

The easiest way to ensure your code is properly formatted is to use pre-commit hooks:

# Install pre-commit
pip install pre-commit

# Install the git hooks
pre-commit install

# Now formatting happens automatically when you commit!
# To run manually on all files:
pre-commit run --all-files

This will automatically format your code with black, isort, and clang-format before each commit.

Python (.py files)

  • black: Code formatting (line length: 100)
  • isort: Import sorting
  • flake8: Style and error checking

Format your Python code manually:

pip install black==24.1.1 flake8==7.0.0 isort==5.13.2
black challenges/ scripts/
isort challenges/ scripts/
flake8 challenges/ scripts/

C++/CUDA (.cu, .cpp, .h files)

  • clang-format: Code formatting (LLVM style with modifications)

Format your C++/CUDA code manually:

sudo apt-get install clang-format
find challenges -name "*.cu" -o -name "*.cpp" -o -name "*.h" | xargs clang-format -i

Mojo (.mojo files)

  • Basic validation checks (file not empty, required imports, @export decorator)

The CI will automatically check all code submissions. Fix any linting errors before submitting your PR.

Difficulty Levels

  • Easy: Single concept, basic kernel launches
  • Medium: Multiple concepts, memory optimizations
  • Hard: Advanced techniques, complex algorithms

Submission Process

  1. Fork the repository
  2. Create a feature branch: git checkout -b challenge/new-challenge-name
  3. Add your challenge with all required files
  4. Submit a pull request with a clear description

Getting Help

  • Open an issue for questions
  • Check existing challenges for examples

Thank you for helping improve LeetGPU!

Contributor Terms

By submitting code, documentation, or any other content to this repository (“Contribution”), you confirm that:

  1. Ownership
    You are the sole author of the Contribution, or have obtained all necessary rights and permissions to grant the licenses below.

  2. Dual License to the Project
    You grant AlphaGPU and its downstream users two non‑exclusive, worldwide, royalty‑free licenses to your Contribution:

    a. Under the repository’s public license,
    Creative Commons Attribution‑NonCommercial‑NoDerivatives 4.0
    (“CC BY‑NC‑ND”), for all public distribution.

    b. Under a commercial‑friendly license to AlphaGPU itself:
    – Grant: AlphaGPU may reproduce, distribute, adapt, sublicense,
    and create derivative works of your Contribution for any purpose,
    including commercial use.
    – No further conditions: AlphaGPU need not attribute further or
    seek additional permission beyond this agreement.

  3. No Additional Restrictions
    You may not impose any further terms on your Contribution that conflict with the licenses granted above.

  4. Pull Request = Acceptance
    Submitting a Pull Request constitutes acceptance of these terms.