The Parallel BMH Algorithm of String Matching

Abstract

One of the critical problems of analyzing internet content is string matching, it is a basic problem in computer fields. During many years study, many classical algorithms were offered. But a bottleneck was met in traditional algorithm: the algorithm’s time complexity of string matching and multiple string matching limits is $O(\frac{n}{m}log_{|\sum|}m)$ and $O(\frac{n}{m}log_{|\sum|}rm)$, many algorithms approach or reach this limit. The article researches parallel algorithm to break the bottlenecks. The paper optimized traditional algorithm by applying multi-core, SSE, AVX instruction. We analyses Suffix based approach, and offer parallel BMH approach, Experimental results show the parallel approach can reach good speedup.