Guanghui Song(宋广辉) bio photo

Guanghui Song(宋广辉)

I am a PhD student at Hunan University, focusing on advanced compilation and high-performance computing.

Email

LinkedIn

Github

Google Scholar

ORCID

Guanghui Song (宋广辉)

“Guanghui” in Chinese means “sheen” in English

I am Guanghui Song, a PhD student at Hunan University under the supervision of Prof. Jie Zhao. I completed my master’s degree at Information Engineering University and then worked as an AI Compiler R&D Engineer at Li Auto Inc., before starting my PhD.

Research Interests

  • Advanced Compilation Techniques: Polyhedral analysis and scheduling, affine loop optimizations, and dataflow-oriented mapping from loop nests and graphs to efficient forms — often through LLVM/MLIR-based passes and reusable compiler pipelines.
  • Hardware Acceleration: Hardware-aware compilation and mapping for NPUs and spatial and reconfigurable architectures (CGRA, FPGA), and HLS-oriented flows—covering operator lowering, pipelining, and resource utilization from IR down to accelerator execution.
  • High Performance Computing: Mixed-precision code generation and autotuning for numerical kernels, lightweight analytical performance models, and scalable execution strategies on parallel and specialized hardware.

My research has appeared in venues such as ISCA, PPoPP, TACO, ASE, and the Journal of Software.

Contact

I’m always interested in collaboration and discussion about my research areas. Feel free to reach out:

News and Updates