Examples

Welcome to the TensorCraft-HPC examples! This section provides hands-on tutorials and practical code samples to help you get started with high-performance AI kernel development.

Quick Links

Example	Description	Difficulty
GEMM Tutorial	Build GEMM from scratch with progressive optimization	🟢 Beginner
FlashAttention	Memory-efficient attention implementation	🟡 Intermediate
Python Bindings	Use TensorCraft from Python	🟢 Beginner

Prerequisites

Before running the examples, make sure you have:

1. CUDA Toolkit 11.0+ installed
2. CMake 3.18+ for building
3. C++17 compatible compiler (GCC 9+, Clang 10+, MSVC 19.28+)
4. Python 3.8+ (optional, for Python bindings)

Running the Examples

C++ Examples

# Clone and build
git clone https://github.com/LessUp/modern-ai-kernels.git
cd modern-ai-kernels

# Build with CUDA support
cmake --preset dev
cmake --build --preset dev

# Run an example
./build/dev/examples/gemm_example

Python Examples

# Install Python package
pip install -e .

# Run Python example
python examples/python/gemm_demo.py

Learning Path

We recommend following this order for the best learning experience:

1. Start with GEMM Tutorial — Learn the fundamentals of CUDA kernel optimization
2. Explore FlashAttention — Understand memory-efficient computing patterns
3. Try Python Bindings — Integrate TensorCraft into your Python workflow

Need Help?

• Check the API Reference for detailed documentation
• Browse Learning Resources for more tutorials
• Open an issue on GitHub if you encounter problems