摘要
设计实现了一种使用统一计算设备架构(CUDA)加速DNA-蛋白质匹配的方法。详细介绍了一种基于退火算法的DNA-蛋白质匹配方法和CUDA的特点,从计算的角度对匹配方法进行了分析。基于CUDA设计实现并行化方法,并根据CUDA的线程调度策略对并行方法进行了优化。实验结果表明,最大可获得15倍左右的加速比。
A method was proposed to accelerate DNA-protein matching, using Compute Unified Device Architecture （CUDA）. Firstly, we introduced an annealing-algorithm-based DNA-protein matching method and the characteristics of CUDA. Then we analyzed the matching method from the computational perspective. Afterwards, we parallelized the matching method using CUDA. Finally, the method is optimized according to the thread scheduling policy of CUDA. Experiments show that up to 15 times speedup can be achieved, in contrast to the CPU-only method.
出处
《电子技术应用》
北大核心
2013年第9期135-138,共4页
Application of Electronic Technique
