基于异构系统的生物序列比对并行处理研究进展  

Advances in Biological Sequence Alignment Parallel Processing Based on Heterogeneous Systems

作　　者：朱香元[1] 李仁发[2] 李肯立[2,3] 胡忠望[1] 

机构地区：[1]肇庆学院计算机学院,肇庆526061 [2]湖南大学信息科学与工程学院,长沙410082 [3]国家超级计算长沙中心,长沙410082

出　　处：《计算机科学》2015年第B11期390-395,399,共7页Computer Science

基　　金：国家自然科学基金重点项目(61133005;61432005);国家自然科学基金青年基金(61402400);广东省自然科学基金(2014A030310288)资助

摘　　要：序列比对工作属于生物信息学的基础性研究领域。由于它具有应用广泛、计算复杂以及海量数据等特点,加之现在高性能计算的兴起,使得近年来序列比对并行处理技术快速发展。首先介绍了序列比对领域高性能计算的新进展,接着从体系结构特征入手对其研究进行分类,并对每类方法的实现细节和性能进行分析比较,从中不难看出访存控制、同步、数据交互以及算法可扩展性等问题均为目前基于异构系统的序列比对并行处理研究的关键点。最后,对该领域的未来研究方向进行了展望。Sequence alignment is a fundamental operation in Bioinformatics. In recent years, there have been extensive studies and rapid progresses in biological sequence alignment parallel processing technologies due to its wide use,high computational complexity and large-scale data. This paper first analyzed the new development of high performance com puting in sequence alignment,and then classified the up-to-date researches based on the applied architectures. Their im- plementation and performance were compared and analyzed in detail. It was pointed out that problems such as memory access control, synchronization, data transferring and scalability of algorithms are the key techniques to the study of biological sequence alignment parallel processing. Finally, some future directions of research were given.

关 键 词：序列比对 并行处理 异构 GPU FPGA Cell BE MIC 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]