高性能计算技术及标准现状分析

doi:10.11896/jsjkx.221100021

摘要/Abstract

摘要： 高性能计算是科技创新体系的重要组成,是知识创新和技术创新的重要能力支撑,是新时期下与理论、实验并重的三大科技创新手段之一。在过去的三十年间,高性能计算取得了以突飞猛进的进展,高性能计算已经进入E级计算时代,我国在高性能计算领域也取得了跨越式的发展,取得了天河、神威、曙光为代表的一系列成果,高性能系统研制水平跻身国际一流行列。随着摩尔定律接近极限,高性能计算技术的性能提升之路面临巨大挑战,在后摩尔时代,将依赖算法、软件和硬件架构去提升高性能计算机系统的终极性能。另一方面,与高性能计算机技术飞速发展相比,高性能计算标准的发展还存在很多不足。文中首先分析了当前国内外高性能计算机技术的发展现状及趋势,然后剖析了当前国内外高性能计算标准的现状及趋势,最后给出了当前发展中国高性能计算机标准的必要性和重要性。

关键词: 高性能计算, E级计算, 计算机体系结构, 标准, 集群, 大规模并行处理机, 后摩尔时代

Abstract: As an indispensable support for knowledge and technological innovation,high-performance computing(HPC) is an important component for scientific and technological innovation system.In the new era,being an alternative way of scientific research,it is equally important as theories and experiments.In the past thirty years,HPC has achieved magic improvement,and has entered the era of exascale computing.China has achieved remarkable development in HPC,and has achieved a series of achievements represented by Tianhe,Sunway,and Dawning.China's high-performance system development level ranks among the top international ranks.Performance gains through semiconductor miniaturization is challenging after Moore's law ends.In the post-Moore era,improvements in computing power,opportunities for growth in computing performance will increasingly come from technologies from software,algorithms,and hardware architecture.Meanwhile,there are still many deficiencies in the development of HPC standards.This paper analyzes the current status and development trends of HPC technology and standards,analyzes the state-of-art of the current HPC standards,and proposes the necessity and importance of developing national HPC standards.

Key words: High-performance computing, Exascale computing, Computer architecture, Standards, Cluster, Massively parallel processor, Post-Moore era

中图分类号:

TP393

陆平静, 熊泽宇, 赖明澈. 高性能计算技术及标准现状分析[J]. 计算机科学, 2023, 50(11): 1-7. https://doi.org/10.11896/jsjkx.221100021

LU Pingjing, XIONG Zeyu, LAI Mingche. Survey on High-performance Computing Technology and Standards[J]. Computer Science, 2023, 50(11): 1-7. https://doi.org/10.11896/jsjkx.221100021

参考文献

[1]LIAO X K,LU K,YANG C Q,et al.Moving from exascale tozettascale computing:challenges and techniques [J].Frontiers of Information Technology and Electronic Engineering,2018,19(10):1236-1244.
[2]ZHENG W.Research trend of large-scale supercomputers and applications from the Top500 and Gordon Bell Prize [J].Science China Information Sciences,2020,63:128-141.
[3]YANG X J,LIAO X K,XU W X,et al.Th-1:first petaflop supercomputer [J].Frontiers of Computer Science in China,2010,4(4):445-455.
[4]WANG R B,LU K,CHEN J,et al.Brief Introduction of TianHeexascale Prototype System [J].Tsinghua Science and Technology,2021,26:361-369.
[5]LIAO X K,PANG Z B,WANG K F,et al.High performance in-terconnect network for TianHe system [J].Journal of Computer Science & Technology,2015,30:259-272.
[6]PANG Z B,XIE M,ZHANG J,et al.The TH Express high-performance interconnect networks [J].Frontiers of Computer Science,2014,8:357-366.
[7]FU H H,LIAO J F,YANG J Z,et al.The Sunway TaihuLight supercomputer:System and applications[J].Science China Information Sciences,2016,59(7):1-16.
[8]GAO J G,ZHENG F,QI F B,et al.Sunway supercomputer ar-chitecture towards exascale computing:Analysis and practice[J].Science China Information Sciences,2021,64:177-197.
[9]YAO W L.High performance computer standards:moving forward in hope and light[J].Information Technology and Standar-dization,2007(6):4-7.
[10]ZENG Y,WANG J.Analysis of High-performance ComputerTechnology and Standards in China [J].Information Technology nd Standards,2016(10):9-12.
[11]The Central Committee of the Communist Party of China and the State Council issued the “National Standardization Development Outline”[EB/OL].(2021-10-10) [2022-11-23].http://www.gov.cn/gongbao/content/2021/content_5647347.
[12]LU K,WANG R B,DONG Y,et al.Challenges and opportunities in the development of exascale high performance computer systems [R].Development Report of 2019－2020.China Computer Science and Technology,2020:418-437.
[13]TAN G M,XUE W,ZHAI J D,et al.Status and trends of high performance computing[R].Development Report of 2018－2019.China Computer Science and Technology,2019.
[14]WANG Y Y,WANG Y W.Micro-nano electronics discipline/industrial development history and rules [J].Science China Information Sciences,2012,42(12):1485-1508.
[15]Thirteen Challenges of Electronic Information Engineering Science and Technology Development(2022) [R].Research on China Electronic Information Engineering Science and Techno-logy Development.2022.
[16]SUN J,LI M,WU H Q,et al.Frontiers and Trends of Micro-electronics in Post Moore Era [J].Bulletin of National Natural Science Foundation of China,2020,34:652-659.
[17]HUANG R,LI M,AN X,et al.New device technology for integrated circuits in the post-Moore era [J].Science China Information Sciences,2012,42,1529-1543.
[18]DENNARD R,GAENSSLEN F.H,RIDEOUT V L,et al.Design of ion-implanted MOSFETs with very small physical dimensions [J].IEEE Journal of Solid-State Circuits,1974,9(5):256-268.
[19]HENNESSY,J L,PATTERSON D A.A New Golden Age for Computer Architecture [J].Communication ACM,2019,62:48-60.
[20]LI B,LU P J.The Evolution of Supercomputer Architecture:aHistorical Perspective [C]//19th CCF Conference on Computer Engineering and Technology.2015:145-153.
[21]JACK D.An overview of high performance computing,the importance of AI/ML and future requirements [R].Invited reports,SMP 2022,Beijing.
[22]JOHN H.The End of the four eras of computer architecture andthe rise of the fifth Era [J].Communication of the Chinese Computer Federation,2021,17(1):38-44.
[23]FEYNMAN R P.There's plenty of room at the bottom [J].Engineering and Science,1960,23:22-36.
[24]MOORE G E.Progress in digital integrated electronics [C]//International Electron Devices Meeting Technical Digest.IEEE,1975:11-13.
[25]CHARLES E L,NEIL C T,JOEL S E,et al.There's plenty of room at the Top:What will drive computer performance after Moore's law? [J].Science,2020,368:6495.
[26]ZENG Y.Comprehensive Review of High-productivity Compu-ter Technologies and Standards [J].Information Technology and Standards,2008,7:17-20.
[27]Founding Conference of New Generation Computing Standards Working Committee Held in Beijing [EB/OL].(2022-08-19) [2022-11-23].http://jl.cesi.cn/cesi/202208/8692.html.
[28]CCF.List of first members of CCF standards committee re-leased.[EB/OL].(2022-04-26) [2022-11-23].https://www.ccf.org.cn/Focus/2022-04-26/761383.shtml.
[29]GJB 7034A-2021,General requirements for military high per-formance cluster computing system [S].Beijing:Equipment Development Department of Central Military Commission,2021.
[30]PCI-SIG.homepage of PCI-SIG[EB/OL].(2022-08-19) [2022-11-23].http://www.pcisig.com.
[31]PCI-SIG.PICMG|Open Standards for Embedded Computing[EB/OL].(2022-08-19) [2022-11-23].http://www.picmg.org/pdf/pcishort.pdf.
[32]PCI-SIG.PCI-Only-e-PCI-X[EB/OL].(2022-08-19) [2022-11-23].http://www.picmg.org/PCI-Only-e-PCI-X.stm.
[33]compactPCI[EB/OL].(2022-04-19) [2022-11-23].https://baike.baidu.com/item/compactPCI.
[34]ATCA Standards and Testing.[EB/OL].(2014-06-20) [2022-11-23].https://yongshengcao.blog.csdn.net/article/.
[35]SJ/T 11721-2018,High performance computing system-BladeServer-Compute Blade Firmware Technical Requirements [S].Beijing:Ministry of Industry and Information Technology,2018.
[36]SJ/T 11719-2018,High performance computing system-BladeServer-Compute Blade Electrical Specifications [S].Beijing:Ministry of Industry and Information Technology,2018.
[37]SJ/T 11720-2018,High performance computing system-BladeServer- Compute Blade mechanical technical requirements [S].Beijing:Ministry of Industry and Information Technology,2018.
[38]SJ/T 11536.1-2015,High performance computing system-Blade Server-Part 1:Management Module Technical Requirements [S].Beijing:Ministry of Industry and Information Technology,2015.
[39]HUGHES J E,SCOLLARD M L,LAND R,et al.BladeCenterprocessor blades,I/O expansion adapters,and units[J].IBM Journal of Research and Development,2005,49(6):837-859.
[40]WUNG D S.Intelligent Platform Management Interface(IPMI) [R].SLAC National Accelerator Laboratory Technical Report,2011.
[41]BHATIA A,MAITY S.Distributed Intelligent Platform Ma-nagement Interface(D-IPMI) System and Method Thereof:U.S.Patent 14/700,843 [P].2015-4-30.
[42]DMTF homepage[EB/OL].[2022-11-23].https://www.dmtf.org/.
[43]Standard on server management:SMASH[EB/OL].(2005-01-12) [2022-11-23].https://blog.csdn.net/iteye_11590/article/details/81313653.
[44]New era data center management standard Redfish[EB/OL].(2021-07-30) [2022-11-23].https://blog.csdn.net/weixin_32821251/article/details/119268494?spm=1001.2014.3001.5501.
[45]SJ/T 11537-2015,High-performance computer cluster monito-ring system technical requirements [S].Beijing:Ministry of Industry and Information Technology,2015.
[46]QX/T 148-2020,Specification for testing and evaluation of high-performance computer systems in the meteorological field [S].Beijing,China Meteorological Bureau,2020.
[47]QX/T 148-2011,Specification for testing and evaluation of high-performance computer systems in the meteorological field [S].Beijing,China Meteorological Bureau,2020.
[48]20213259-T-469,Information technology－High performancecomputing system-Technical requirement for management and monitor platform [S].Beijing:Ministry of Industry and Information Technology,2021.
[49]20190820-T-469,Test methods of energy efficiency for high performance computer system [S].Beijing:Ministry of Industry and Information Technology,2019.
[50]InfiniBand Networking Solutions [EB/OL].http://network.nvidia.com/en-us/networking/infiniband-switching.
[51]InfiniBand Trade Association.InfinibandTM Architecture Specification;Release 1.2.1 [S].Infiniband Trade Association:Beaverton,OR,USA,2007;Volume 1.
[52]BIRRITTELLA M S,DEBBAGE M,HUGGAHALLI R,et al.Enabling Scalable High-Performance Systems with the Intel Omni-Path Architecture [J].IEEE Micro,2016,36(4):38-47.
[53]TRADER T.With New Owner and New Roadmap,and Inde-pendent Omni-Path Is Staging a Comeback.HPC Wire [EB/OL].https://www.hpcwire.com/2021/07/23/with-new-ow-ner-and-new-roadmap-an-independent-omni-path-is-staging-a-come-back/.
[54]MURPHY P.Cornelis networks omni-path:Purpose built high-performance fabrics for HPC/HPDA/AI [C]//Proceedings of the Supercomputing Frontiers Europe.2021.
[55]DE S D,DI G S,MCMAHON K H,et al.An in-depth analysis of the slingshot interconnect [C]//Proceedings of the International Conference for High Performance Computing,Networking,Sto-rage and Analysis.2020:1-14.
[56]HPE.HPE Slingshot_The Interconnect for the Exascale EraTechnical White Paper [EB/OL].https://assets.ext.hpe.com/is/content/hpedam/documents/a50002000-2999/a50002368/a50002368enw.pdf.
[57]AJIMA Y,INOUE T,HIRAMOTO S,et al.Tofu interconnect 2:System-on-chip integration of high-performance interconnect [C]//International Conference on Supercomputing.2014:498-507.
[58]AJIMA Y,KAWASHIWA T,OKAMOTO T.The tofu interconnect D [C]//Proceedings of the IEEE International Confe-rence on Cluster Computing.2018:646-654.
[59]SAID D,THIBAUT P S,PANZIERA J P,et al.The BXI interconnect architecture [C]//Proceedings of the IEEE 23rd Annual Symposium on High-Performance Interconnects.2015:18-25.
[60]GJB 5615-2006,Design requirement for interconnect network of military supercomputer [S].Beijing:Equipment Development Department of Central Military Commission,2006.
[61]GJB 5614-2006,Specification for memory system of military supercomputer [S].Beijing:Equipment Development Department of Central Military Commission,2006.
[62]GJB 5169-2004,General specification for disk array of high performance computer [S].Beijing:Equipment Development Department of Central Military Commission,2004.
[63]GJB 4456-2002,Design requirement for parallel language andcompilation system of military supercomputer [S].Beijing:Equipment Development Department of Central Military Commission,2002.
[64]GJB/Z 4894-2018,Design guide for parallel program and development environment of supercomputer [S].Beijing:Equipment Development Department of Central Military Commission,2018.
[65]GJB 4894-2003,Design requirement for parallel program development environment of supercomputer [S].Beijing:Equipment Development Department of Central Military Commission,2003.
[66]GJB 4893A-2018,Design requirement for parallel operating system of supercomputer [S].Beijing:Equipment Development Department of Central Military Commission,2018.
[67]GJB 4893-2003,Design requirement for parallel operating system of supercomputer [S].Beijing:Equipment Development Department of Central Military Commission,2003.
[68]BERNHOLDT D E,BOEHM S,BOSILCA G,et al.A survey of MPI usage in the US exascale computing project [J].Concurrency & Computation:Practice & Experience,32(3):e4851.
[69] CHUNDURI S,PARKER S,BALAJI P,et al.Characterization of MPI usage on a production supercomputer [C]//Proceedings of International Conference for High Performance Computing,Networking,Storage,and Analysis.IEEE Press,2018,30.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed