摘要
新疆天文台Taurus高性能计算系统由1个登陆节点、16个计算节点、2个I/O节点、100 TB高速存储组成。其中央处理器理论双精浮点计算能力6.758 4 Tflops,通过Linpack测试实际峰值为6.289 Tflops,可提供计算能力为理论值的93.06%;图形处理器理论计算能力18.72 Tflops,实际测试计算峰值为14.882 Tflops,计算效率为理论值的79.5%。系统计算与存储节点均采用56 Gb Infiniband交换互连,通过IOZone测试存储系统,单节点写速度可达460 MB/s,多节点写速度可达800 MB/s。系统已在多相滤波及消干扰图形处理器算法加速、蒙特卡罗模拟等领域得到了应用。
Taurus high performance computing system of Xinjiang Astronomical Observatory has 1 login node,16 compute nodes,2 I/O nodes and 100TB high-speed storage.In theory,the double precision floating-point computation capacity of CPUs is 6.7584Tflops.The actual peak turns out to be 6.289Tflops tested by Linpack,the available computation capability is 93.06%of the theoretical value.The computation capacity of the GPUs is 18.72Tflops in theory,while its practical peak is 14.882Tflops,the available computation capability is 79.5%of the theoretical value.The calculation nodes and the storage nodes are connected by 56Gb Infiniband network.Using IOZone for testing the storage performance,single-node writing reaches 460MB/s and multi-node writing can be 800MB/s.The Taurus HPC system has been applied in various fields such as GPU algorithm acceleration and Monte Carlo simulation.
作者
张海龙
冶鑫晨
王杰
聂俊
陈龙飞
托乎提努尔
崔辰州
李长华
朱艳
张萌
Zhang Hailong;Ye Xinchen;Wang Jie;Nie Jun;Chen Longfei;Tohtonur;Cui Chenzhou;Li Changhua;Zhu Yan;Zhang Meng(Xinjiang Astronomical Observatory.Chinese Academy of Sciences.Urumqi 830011.China;Key Laboratory of Radio Astronomy.Chinese Academy of Sciences.Nanjing 210008.China;National Astronomical Observatories.Chinese Academy of Sciences.Beijing 100101.China;University of Chinese Academy of Sciences.Beijing 100049.China)
出处
《天文研究与技术》
CSCD
2018年第3期332-339,共8页
Astronomical Research & Technology
基金
国家重点基础研究发展计划(973计划)(2015CB857100)
国家自然科学基金(11503075,U1531125)
中国科学院青年创新促进会
中国科学院天文台站设备更新及重大仪器设备运行专项经费资助