Mini-Inference Engine

English | 简体中文 | Docs

Mini-Inference Engine is a CUDA GEMM optimization learning project that packages progressive matrix multiplication kernels, a lightweight inference runtime, and profiling-oriented experimentation into one repository.

Repository Overview

• CUDA kernels and runtime headers in src/ and include/
• Technical docs under docs/
• Benchmarks and demos in benchmarks/
• GitHub Pages site for documentation entry, reading paths, and project updates

Quick Start

cmake --preset default
cmake --build --preset default
ctest --preset default

For an optimized build without tests:

cmake --preset release
cmake --build --preset release
./build-release/benchmark

Docs

• Project docs: https://lessup.github.io/mini-inference-engine/
• Site home explains what to read first for architecture, optimization, and API details
• See CONTRIBUTING.md for contribution workflow

License

MIT License