Parallel Community Detection Algorithm Using a Data Partitioning Strategy with Pairwise Subdomain Duplication

Diana Palsetia, William Hendrix, Sunwoo Lee, Ankit Agrawal, Wei-keng Liao, Alok Choudhary

ISC, 2016

This paper presents PMEP, a distributed-memory parallel community detection algorithm that partitions a graph into subgraphs and assigns each subgraph pair to a process (trading some duplicated computation for reduced communication), then runs the MEP algorithm per pair, achieving speedups of 126.95x on 190 MPI processes (synthetic data) and 204.22x on 1225 processes (real-world data) without sacrificing result quality.

BibTeX

@inproceedings{palsetia2016parallel,
  title={Parallel community detection algorithm using a data partitioning strategy with pairwise subdomain duplication},
  author={Palsetia, Diana and Hendrix, William and Lee, Sunwoo and Agrawal, Ankit and Liao, Wei-keng and Choudhary, Alok},
  booktitle={International Conference on High Performance Computing},
  pages={98--115},
  year={2016},
  organization={Springer}
}