Title
Orchestrating parallel detection of strongly connected components on GPUs.
Abstract
•Analysis on graph properties: skewed SCC sizes and dynamically changing structure.•A hybrid parallelism method is proposed to deal with skewed SCC sizes.•Traversal strategies are customized to fit dynamically changing graph structure.•Our method outperforms existing GPU and OpenMP implementations.
Year
DOI
Venue
2018
10.1016/j.parco.2017.11.001
Parallel Computing
Keywords
Field
DocType
Strongly connected components,GPU,Real-world graphs,Hybrid parallelism
Central processing unit,Graph traversal,Graph property,Computer science,Parallel algorithm,Parallel computing,Exploit,Theoretical computer science,Implementation,Strongly connected component,Speedup
Journal
Volume
ISSN
Citations 
78
0167-8191
3
PageRank 
References 
Authors
0.40
23
7
Name
Order
Citations
PageRank
Xuhao Chen1407.43
cheng chen271.12
Jie Shen31018.05
Jianbin Fang426525.31
Tao Tang5427.44
Canqun Yang618829.39
Zhi-Ying Wang7870127.04