I'm a Research Software Engineer working on high-performance scientific software, HPC, and GPU computing.
My work focuses on developing and optimising numerical software for heterogeneous CPU/GPU systems using modern C++, CUDA, and OpenMP. My application background is primarily in Computational Fluid Dynamics (CFD), with an emphasis on numerical algorithms, performance engineering, and maintainable scientific software.
- Programming: C++, CUDA, Fortran, Python, OpenMP
- HPC: CPU/GPU parallel computing, heterogeneous computing, performance optimisation
- Scientific Computing: Numerical methods, CFD, finite volume methods
- Performance Engineering: GPU profiling, benchmarking, kernel optimisation
- Software Engineering: CMake, Git, Linux, Docker, CI, automated testing
High-performance C++ CFD library for heterogeneous computing
NeoN is an open-source C++ library providing data structures, parallel algorithms, and numerical infrastructure for finite-volume CFD on modern heterogeneous computing systems.
My contributions include:
- Improving solver infrastructure and maintaining consistent interfaces and data types
- Implementing uniform mesh-generation functionality and associated tests
- Establishing automated testing across heterogeneous NVIDIA and AMD GPU environments
- Improving project documentation
Performance-portable finite-volume CFD framework
NeoFOAM provides finite-volume CFD solvers designed for execution on heterogeneous CPU and GPU systems.
My contributions include:
- Integrating GPU profiling and performance-analysis tools to guide optimisation
- Profiling scientific workloads on GPU-based HPC systems to identify performance bottlenecks
- Establishing automated testing across heterogeneous NVIDIA and AMD GPU environments
- Improving project documentation
- High-performance scientific computing
- Scientific software engineering
- GPU computing and performance optimisation
- Numerical algorithms and data structures
- CFD and finite-volume methods
- Modern C++ for HPC
Email: wangchihta@gmail.com