Using GPUs for Graph Partitioning via Embedding

The project is about using and enhancing an existing Graph Embedding tool for Graph Partitioning. Both of these problems are important in practice and various tools have been developed. Although they are related, the relationship has only been briefly investigated. In this project, the students will use our in-house embedding tool, GOSH (https://ieeexplore.ieee.org/document/9623416), to partition the graphs into given number of parts. The tool is written in CUDA. Hence, CUDA knowledge for GPU programming is a must.