基于GPU的非相干消色散算法  

Incoherent Dedispersion Algorithm Based on GPU

作　　者：托乎提努尔[1,3] 张海龙 王杰[1,3] 冶鑫晨 Toktonur;Zhang Hailong;Wang Jie;Ye Xinchen(Xinjiang Astronomical Observatory,Chinese Academy of Sciences,Urumqi 830011,China;Key Laboratory of Radio Astronomy,Chinese Academy of Sciences,Nanjing 210008,China;National Astronomical Data Center,Beijing 100101,China)

机构地区：[1]中国科学院新疆天文台,新疆乌鲁木齐830011 [2]中国科学院射电天文重点实验室,江苏南京210008 [3]国家天文科学数据中心,北京100101

出　　处：《天文研究与技术》2021年第1期60-68,共9页Astronomical Research & Technology

基　　金：国家自然科学基金(11873082,11803080);国家重点研发计划(2018YFA0404704);中国科学院青年创新促进会;“西部之光”(2019-XBQNXZ-B-018);国家天文科学数据中心;中国科学院科学数据中心体系资助.

摘　　要：针对脉冲星信号实时消色散处理的需求,实现了基于图形处理器(Graphics Processing Unit,GPU)的非相干消色散算法。采用高性能并行计算方法对非相干消色散算法的多线程处理进行了深入研究,提出了算法的并行化加速方案,解决了消色散算法计算量大、无法实时处理的问题。分析算法的密集型计算部分,高效利用图形处理器的层次存储结构,提高了图形处理器的资源利用率,减少了计算时间,显著提升了非相干消色散算法的计算性能。In order to meet the requirements of real-time de-dispersion processing of pulsar signals,a GPU based incoherent de-dispersion algorithm is implemented.We use high-performance parallel computing methods to conduct in-depth research on the incoherent de-dispersion algorithm for multi-thread processing,and propose a parallel acceleration scheme for the algorithm,which solves the problem that the de-dispersion algorithm has a large amount of calculation and cannot be processed in real time.We analyze the intensive computing parts of the algorithm,efficiently use the hierarchical storage structure of the GPU,improve the utilization rate of GPU computing resources,thereby reducing the calculation time,and significantly improving the calculation performance of the incoherent de-dispersion algorithm.

关 键 词：消色散 图形处理器 并行计算 实时处理 

分 类 号：P111.44[天文地球—天文学]