国产万亿次机群系统NPB性能测试分析

对3个国产万亿次机群系统进行了NPB性能测试分析，重点研究大规模并行处理时(处理器数目达到上千个)的性能特点和趋势．分析了不同的处理器、互连网络等系统配置对NPB性能的影响，发现NPB的8个程序在3个万亿次机器上的性能特点和表现并不一致，表明国产高性能机群在设计上正在逐渐走出同质化的趋势，向多样化发展．进一步分析表明，目前NPB程序的可扩展性可以达到几百个处理器，但尚不能达到上千个处理器，NPB程序能发挥出的系统峰值的百分比仍然徘徊在10％左右，机群系统的并行可扩展性和应用程序对机器运算潜能的利用还需要进一步提高．对于处理器数目达到上千个的万亿次机群系统来说，对集合通信和细粒度通信能力的支持亟需提高．

Performance Analysis of NPB Benchmark on Domestic Tera-Scale Cluster Systems

In this paper, NPB benchmarking is performed on three domestic tera-scale cluster systems with emphasis on the performance characteristics and trends when carrying out tera-scale parallel computing on systems with thousands of processors. The effects of different system configurations (processor, interconnection network, etc.) on the final NPB performance are analyzed and it is found that the programs in NPB suites got their best performance on different clusters. Through further analysis, it is indicated that the scalability of NPB programs can reach hundreds of processors, but can't reach thousands of processors. Most of the NPB programs can only exploit around 10% of the system peak performance, so the scalability of cluster systems and real application performance on tera-scale cluster systems need further improvement. For manufacturing of tera-scale cluster systems with thousands of processors, the performance of collective communication and fine-grained message passing needs further improvement.