摘要
ncRNA(非编码RNA)是一类重要的遗传物质,它通过多种机制调控着基因的表达。由于缺少编码RNA基因所具有的典型特征,ncRNA基因的检测成为生物信息学RNA研究领域的热点问题。QRNA是目前该领域最典型使用最广泛的程序之一,但受限于O(L3)计算复杂度,传统的软件预测方法并不能满足日常研究的需要。本文基于FPGA平台实现了一种细粒度的并行ncRNA检测算法,利用CPU加FPGA的方案对QRNA程序实现细粒度并行,采用按矩阵列循环划分的任务分配策略实现处理单元间的负载平衡;采用数据预取、滑动窗口和数据传递流水线实现处理单元间的数据重用,减少片外访存开销。在单片FPGA上集成了由8个处理单元构成的计算阵列。实验结果表明,与运行在AMD四核9650处理器上的QRNA-2.0.3c程序相比,可获得超过18倍的加速效果,并且FPGA加速器功耗仅为通用微处理器平均功耗的20%。
Non-coding RNAs have important functional roles in biological processes and become a central research interest in modern molecular biology.QRNA is one of the powerful programs and has been widely used as an efficient analysis tool to detect the ncRNA gene at present.Unfortunately,the O（L3） computing requirements and complicated data dependency greatly limit the usefulness of the QRNA package with the explosion in gene databases.In this paper,we present a fine-grained parallel QRNA prototype system for accelerating the ncRNA gene detection application on the FPGA chip.The experimental results show a factor of more than 18x speedup over the QRNA-2.0.3c software running on a PC platform with the AMD Phenom 9650 Quad CPUs for pairwise sequence alignments with 996 bps,however the power consumption of our FPGA accelerator is only about 20% of the of the latter.
出处
《计算机工程与科学》
CSCD
北大核心
2011年第12期153-158,共6页
Computer Engineering & Science
基金
国家863计划资助项目(2007AA01Z106)
