Projects

Boyle — High-Performance C++ Numerical Computing Library

The Mathematical Foundation for Autonomous Vehicles and Robotics

• Implemented entirely in C++23 with 20,000+ lines of code, extensively leveraging modern template metaprogramming
• Primarily header-based library with fine-grained CMake targets — link only what you need
• Core Modules:
  • boyle::math — Dense/sparse linear algebra, piecewise curves (linear/cubic/quintic splines), FFT, Chebyshev polynomials, SIMD vectorization (AVX-512)
  • boyle::cvxopm — Convex optimization: QP solver (OSQP), L-BFGS, BFGS, Nelder-Mead, line search
  • boyle::kinetics — Motion planning: route/path/motion modeling, cubic/quintic acceleration & offset models via QP optimization
  • boyle::common — FSM, logging (spdlog-based), aligned allocators, RAII utilities
• Dual build systems (CMake + xmake), cross-platform (Linux/macOS/Windows)
• Fully automated GitHub Actions CI/CD with comprehensive unit test coverage
• License: BSD-3-Clause

---

LhcVaspTools — VASP Data Analysis Toolkit

Python toolkit for reading data from VASP output files (vaspout.h5) and creating publication-quality plots. Developed during research at MPI-CPFS for post-processing DFT calculation results.

---

Open Source Contributions

• Arch Linux — Fixed bcachefs installation image issue
• xdit — Ported distributed inference framework to Moore Threads GPU
• long-context-attention — Ported long-context attention mechanism to Moore Threads GPU