基于GPU集群的空间VLBI射电源条纹搜索研究  

Research on Space VLBI Radio Source Fringe Search Based on GPU Cluster

作　　者：段学铭 童力[1,3,5] 郑为民 张娟[1,3,4,5] DUAN Xue-ming;TONG Li;ZHENG Wei-min;ZHANG Juan(Shanghai Astronomical Observatory,Chinese Academy of Sciences,Shanghai 200030,China;University of Chinese Academy of Sciences,Beijing 100049,China;Key Laboratory of Radio Astronomy,Chinese Academy of Sciences,Beijing 100101,China;National Public Science Data Center for Basic Disciplines,Beijing 100190,China;Shanghai Key Laboratory of Space Navigation and Positioning Techniques,Shanghai 200030,China)

机构地区：[1]中国科学院上海天文台,上海200030 [2]中国科学院大学,北京100049 [3]中国科学院射电天文重点实验室,北京100101 [4]国家基础学科公共科学数据中心,北京100190 [5]上海市空间导航与定位技术重点实验室,上海200030

出　　处：《天文学进展》2024年第3期494-505,共12页Progress In Astronomy

基　　金：国家自然科学基金(11973011);国家基础科学数据共享服务平台(U1931135)。

摘　　要：探月工程四期将发射嫦娥七号“鹊桥二号”中继星,搭载4.2 m口径的抛物面望远镜,与地面射电望远镜构成首个月轨空间VLBI。空间VLBI受中继星轨道扰动和星载设备时延等因素影响,射电源的预报时延模型无法引导相关处理机正常工作,需通过基于实测信号条纹搜索的方法找到符合要求的高精度时延模型。为此,提出一种基于GPU集群的空间VLBI射电源条纹搜索算法,研究在GPU集群上负载均衡的任务分配方法,并行实现相关处理模块和残余值搜索模块。通过RadioAstron 1 min观测数据验证,相较于目前CPU集群平台,GPU集群处理速度提升了27.0倍。The fourth phase of the Lunar Exploration Project will launch the Chang’e-7 relay satellite,carrying a 4.2-meter parabolic telescope.Together with ground-based radio telescope,it will form the first Lunar Orbit Space VLBI system.Space VLBI is influenced by the relay satellite’s orbit perturbations and onboard equipment delays,and the predicted time delay model for radio source may not effectively guide the normal operation of the correlator.It is necessary to employ a fringe search based on actual measurement signals to find a high-precision time delay model that meets the requirements.Currently,fringe searches are accelerated using CPU cluster,but the search time is prolonged,lasting up to 2700s,which fails to meet the practical engineering requirements.A space VLBI radio source fringe search algorithm based on GPU cluster is proposed.The research involves studying task distribution for load balancing on GPU cluster and parallel implementation of correlation modules and residual value search modules.Validated with RadioAstron observation data,the processing speed of GPU cluster has improved about 27.0 times compared to CPU cluster.The research outcomes can be applied to subsequent space VLBI projects.

关 键 词：空间VLBI 条纹搜索 GPU集群 相关处理 

分 类 号：P111.44[天文地球—天文学]