Parallel Suffix Array Construction Algorithm for GPU Computing

Suffix array is widely used in sequence analysis,string matching and text compression,over the last few years,suffix array algorithms for construction and application have constituted an intense area of research within computer science.After the analysis and comparison of existing serial algorithms,we propose a new compact parallel prefix-doubling suffix array construction algorithm fit for GPU computing by replacing group with sort.The algorithm can run independently for parallel suffix array construction,and it is also compatible and cooperative with existing prefix-doubling algorithms for peak performance.The experiment results show that the algorithm is simple,fast and scalable for real-world data processing,and especially efficient for small alphabet data.