摘要
cmsearch程序是目前最流行的同源RNA搜索工具之一,其最大的缺陷在于搜索速度过慢,严重影响了其应用范围。文章基于序列划分策略采用两级工作池方式实现同源RNA的并行搜索,并对其中所涉及的通信和负载平衡问题进行了优化。测试结果表明,此文方法具有良好的可扩展性,在120个处理器时,并行效率可达86.3%,能够用于全基因组范围内的大规模同源RNA序列搜索。
Cmsearch program is one of the most popular tools for searching homologous RNA. However, its speed is too slow, which seriously limits its application. In this paper, we present a method for parallel searching which employs sequence partition strategy and two-level work pool approach. Then, we optimized the communication and load balance involved in our method. As experiments shown, our parallel method has good scalability, which achieves 86.3% efficiency at 120 CPUs, and can be used in solving large-scale homologous RNA search problem in genomic region.
出处
《微电子学与计算机》
CSCD
北大核心
2006年第9期1-3,9,共4页
Microelectronics & Computer
关键词
同源RNA序列
两级工作池
并行化
Homologous RNA sequence, Two-level work pool, Parallelization
