继理论科学和实验科学之后，高性能计算成为人类科学研究的第三大范式。作为科技创新的重要手段，高性能计算广泛应用于核爆模拟、天气预报、工程计算等众多领域，是当代科技竞争的战略制高点，集中体现一个国家的综合实力。本文介绍高性能计算发展的历史和现状，分析当前高性能计算所面临的问题和挑战，探讨高性能计算未来的发展方向。
Following theoretical science and experimental science, high-performance computing has become the third largest paradigm of human exploration. As an important means of technology innovation, high-performance computing is widely used in many areas, such as nuclear explosion simulation, climate simulation, and engineering calculations. Moreover, it is a strategic commanding point of modern scitech competition and embodies the comprehensive strength of a nation. In this paper, we describe the history and current situation of highperformance computing, then analyzes the problems and challenges. Also, we discuss the future direction of high-performance computing.
[1] Higbie L C. Tutorial: Supercomputer architecture[J]. Computer, 1973, 6(12): 48-58.
[2] IDC. 2015, IDC HPC Update at ISC'15[EB/OL]. [2016-04-28]. http://www.slideshare.net/insideHPC/hpc-market-update-from-idc-51061896.
[3] Solomon S. Climate change 2007-the physical science basis: Working group I contribution to the fourth assessment report of the IPCC[M]. Cambridge:Cambridge University Press, 2007.
[4] Kohl S, Leitzl K H, Schmidt M. Transient numerical simulation of CO2 laser fusion cutting of metal sheets: Simulation model and process dynamics[C]//Proceedings of the 37th MATADOR Conference. Manchester: Springer, 2012: 403.
[5] Mohamed A, Yu D, Deng L. Investigation of full-sequence training of deep belief networks for speech recognition[C]. INTERSPEECH 2010, 11th Annual Conference of the International Speech Communication Association, Makuhari, Chiba, Japan, September 26-30, 2010.
[6] Seide F, Li G, Yu D. Conversational speech transcription using context-dependent deep neural networks[C]. INTERSPEECH 2011, 12th Annual Conference of the International Speech Communication Association, Florence, Italy, August 27-31, 2011.
[7] The New York Times. How many computers to identify a cat? 16,000[J/OL]. Communications of the Acm, 2012-06-26[2016-04-28], http://www.cs.cornell.edu/courses/CS6700/2013sp/readings/04-a-Deep-Learning-NYT.pdf.
[8] Chi K R. Neural modelling: Abstractions of the mind[J]. Nature, 2016, 531(7592): S16-S17.
[9] Moore G. Moore's law[J]. Electronics Magazine, 1965, 38(8): 114.
[10] The Top500 List 2015. [EB/OL]. [2016-04-20]. Available: http://www.Top500.org.
[11] Bryant R E, Polk W T. The National Strategic Computing Initiative[C]//Proceedings of the International Conference for High Performance Computing, Net-working, Storage and Analysis: ACM, Austin 2015.
[12] Murphy R C, Wheeler K B, Barrett B W, et al. Introducing the graph 500[J]. Cray User's Group (CUG), 2010, 14(3): 15555-15558.
[13] The Green500 List. [EB/OL]. [2016-04-22]. Available: http://www.Green500.org.
