会议专题

Implementation of String Match Algorithm BMH on GPU

String, match algorithm is widely used in the area of data mining. In this paper, we present an approach for elevating the performance of this algorithm via GPU (Graphic Processing Unit). With the rapid development of Graphics Processing Unit to many-core multiprocessors, it shows great potential in many applications and high performance computing. Especially, the heterogeneous architecture CPU+GPU shows enthusiastic capacity to accelerate parallel applications. Till now, some research has been done for parallel implementation of string match algorithm on GPU. But there are some constraints for acceleration of this algorithm: firstly, if-brancb exists in the algorithm which can cripple the performance of the parallel implementation; secondly, improper memory access may bring heavy latency generated by bankiconflict This paper presents the optimization of global memory access, shared memory access and the elimination of if-branch for elevating the performance of string match algorithm BMH.

BMH string match parallel GPU CUDA

Jun Rui Zhou Hong An Xiao Mei Li Wei Zhou Min Xu

Department of Computer Science and Technology, University of Science and Technology of China Advanced Computer System and Architecture Laboratory Hefei, Anhui, China

国际会议

2011 International Conference on Database and Data Mining(ICDDM 2011)(2011年数据库和数据挖掘国际会议)

三亚

英文

181-185

2011-03-25(万方平台首次上网日期,不代表论文的发表时间)