Status, Challenges, and Trends of Data-Intensive Supercomputing

doi:10.11871/jfdc.issn.2096-742X.2023.03.006

Abstract

Abstract:

[Objective] This paper is to provide a comprehensive and systematic overview of the development history, mainstream system architecture, typical applications, and computation and storage subsystems of data-intensive supercomputing, point out the future development trend, and provide references for further data-intensive supercomputing optimization. [Methods] This paper first sorts out the key concepts of data-intensive supercomputing and analyzes the support to the data-intensive applications by existing platforms. Then the real demand for data-intensive applications from the mainstream academic and industrial communities are illustrated. Finally, the future trends and potential challenges of data-intensive supercomputing are discussed and a corresponding supercomputing system evaluation model is developed. [Results] Relevant researchers and practitioners can quickly understand the key concepts and development status of supercomputing technology from this paper, and precisely capture the current and future data-intensive supercomputing research hotspots and key problems that need to be solved. [Conclusions] The problems such as the optimization on complex data type and mixed workload, and multi-protocol support and interoperability which are faced by the data-intensive supercomputing storage systems will become hot research and development issues in the coming years.

Key words: data-intensive supercomputing, I/O intensive supercomputing, high performance data analytics, parallel processing system, supercomputing storage system

WEI Jia, CHEN Mo, WANG Longxiang, REN Pei, LEI Yujia, QU Yuqi, JIANG Qiyu, DONG Xiaoshe, WU Weiguo, ZHANG Kaili, ZHANG Xingjun. Status, Challenges, and Trends of Data-Intensive Supercomputing[J]. Frontiers of Data and Computing, 2023, 5(3): 66-91, https://cstr.cn/32002.14.jfdc.CN10-1649/TP.2023.03.006.

Figures/Tables 21

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Fig.6

Fig.7

Fig.8

Fig.9

Fig.10

Fig.11

Fig.12

Table 1

Table 2

Table 3

Table 4

Fig.13

Fig.14

Fig.15

Fig.16

Table 5

References 86

[1]	Middleton A M. Data Intensive Supercomputing Solut-ions[M]//Big Data Technologies and Applications. Cham: Springer-Verlag, 2016:257-306.
[2]	Kleppmann M. Designing data-intensive applications: The big ideas behind reliable, scalable, and maintainable systems[M]. Sebastopol: O’Reilly Media Inc, 2017:15-16.
[3]	Harchol-Balter M. Performance modeling and design of computer systems: queueing theory in action[M]. Cambr-idge: Cambridge University Press, 2013:110-119.
[4]	Hager G, Wellein G. Introduction to high performance computing for scientists and engineers[M]. Boca Raton: CRC Press, 2010:19-21.
[5]	Schmidt B, Hildebrandt A. Next-generation sequencing: big data meets high performance computing[J]. Drug discovery today, 2017, 22(4): 712-717. doi: S1359-6446(17)30058-2 pmid: 28163155
[6]	Ding Z, Yang B, Chi Y, et al. Enabling smart transport-ation systems: A parallel spatio-temporal database appro-ach[J]. IEEE Transactions on Computers, 2015, 65(5): 1377-1391. doi: 10.1109/TC.2015.2479596
[7]	Henz B J, Elliot L, Barton M, et al. High-Performance Computing for the Next Generation Combat Vehicle[R]. Maryland: US Army Research Laboratory, 2018.
[8]	Puertas-Martín S, Banegas-Luna A J, Paredes-Ramos M, et al. Is high performance computing a requirement for novel drug discovery and how will this impact academic efforts?[J]. Expert opinion on drug discovery, 2020, 15 (9): 981-985. doi: 10.1080/17460441.2020.1758664 pmid: 32345062
[9]	Orhan A E. Robustness properties of Facebook’s ResN-eXt WSL models[J/OL]. arXiv preprint arXiv: 1907.07640, 2019-12-10 [2022-05-07]. https://arxiv.org/abs/1907.07640.
[10]	Rajak R. A comparative study: Taxonomy of high perfo-rmance computing (HPC)[J]. International Journal of Electrical and Computer Engineering, 2018, 8(5): 3386.
[11]	Shimizu T. Supercomputer Fugaku: Co-designed with application developers/researchers[C]. The 2020 IEEE Asian Solid-State Circuits Conference (A-SSCC), Man-hattan: IEEE, 2020: 1-4.
[12]	Gao J, Zheng F, Qi F, et al. Sunway supercomputer arc-hitecture towards exascale computing: analysis and prac-tice[J]. Science China Information Sciences, 2021, 64(4): 1-21.
[13]	Usman S, Mehmood R, Katib I. Big data and HPC conve-rgence for smart infrastructures: a review and proposed architecture[M]. Switzerland: Springer Cham, 2020: 561-586.
[14]	Kodama Y, Odajima T, Arima E, et al. Evaluation of power management control on the supercomputer fugaku[C]. The 2020 IEEE International Conference on Cluster Computing (CLUSTER), Manhattan: IEEE, 2020: 484-493.
[15]	Nakao M, Ueno K, Fujisawa K, et al. Performance eval-uation of supercomputer fugaku using breadth-first search benchmark in Graph500[C]. The 2020 IEEE International Conference on Cluster Computing (CLUSTER), Manh-attan: IEEE, 2020: 408-409.
[16]	Kudo S, Nitadori K, Ina T, et al. Implementation and numerical techniques for one EFlop/s HPL-AI benchmark on Fugaku[C]. The 2020 IEEE/ACM 11th Workshop on Latest Advances in Scalable Algorithms for Large-Scale Systems (ScalA), Manhattan: IEEE, 2020: 69-76.
[17]	Odajima T, Kodama Y, Tsuji M, et al. Preliminary per-formance evaluation of the fujitsu a64fx using hpc appl-ications[C]. The 2020 IEEE International Conference on Cluster Computing (CLUSTER), Manhattan: IEEE, 2020: 523-530.
[18]	Zhang K, Su H, Zhang P, et al. Optimization and Perf-ormance Modeling of Stencil Computations on ARM Architectures[C]. 2020 IEEE 22nd International Confer-ence on High Performance Computing and Communic-ations; IEEE 18th International Conference on Smart City; IEEE 6th International Conference on Data Science and Systems(HPCC/SmartCity/DSS), Manhattan: IEEE, 2020: 113-121.
[19]	Kahle J A, Moreno J, Dreps D. 2.1 Summit and Sierra: designing AI/HPC supercomputers[C]. 2019 IEEE Inter-national Solid-State Circuits Conference-(ISSCC), Man-hattan: IEEE, 2019: 42-43.
[20]	Wang R, Tobar R, Dolensky M, et al. Processing full-scale square kilometre array data on the summit superco-mputer[C]. SC20:International Conference for High Performance Computing, Networking, Storage and Anal-ysis, Manhattan: IEEE, 2020: 1-12.
[21]	Luo L, Straatsma T P, Suarez L E A, et al. Pre-exascale accelerated application development: The ORNL Summit experience[J]. IBM Journal of Research and Develo-pment, 2020, 64(4): 11.
[22]	Hernández B, Somnath S, Yin J, et al. Performance eval-uation of python based data analytics frameworks in summit: Early experiences[C]. Smoky Mountains Com-putational Sciences and Engineering Conference, Cham: Springer, 2020: 366-380.
[23]	Womble D E, Shankar M, Joubert W, et al. Early expe-riences on summit: Data analytics and AI applications[J]. IBM Journal of Research and Develop-ment, 2019, 63(6): 1-9.
[24]	Fu H, Liao J, Yang J, et al. The Sunway TaihuLight sup-ercomputer: system and applications[J]. Science China Information Sciences, 2016, 59(7): 1-16.
[25]	Chen Q, Chen K, Chen Z N, et al. Lessons learned from optimizing the Sunway storage system for higher applic-ation I/O performance[J]. Journal of Computer Science and Technology, 2020, 35(1): 47-60. doi: 10.1007/s11390-020-9798-5
[26]	Lu Y T, Cheng P, Chen Z G. Design and Implementation of the Tianhe-2 Data Storage and Management System[J]. Journal of Computer Science and Technology, 2020, 35(1): 27-46. doi: 10.1007/s11390-020-9799-4
[27]	臧大伟, 曹政, 孙凝晖. 高性能计算的发展[J]. 科技导报, 2016, 34(14): 22-28.
[28]	Lin H, Tang X, Yu B, et al. Scalable graph traversal on sunway taihulight with ten million cores[C]. The 2017 IEEE International Parallel and Distributed Processing Symposium (IPDPS), Manhattan: IEEE, 2017: 635-645.
[29]	Dong W, Li K, Kang L, et al. Implementing molecular dynamics simulation on the Sunway TaihuLight system with heterogeneous many-core processors[J]. Concur-rency and Computation: Practice and Experience, 2018, 30(16): e4468.
[30]	Seal S K, Lim S H, Wang D, et al. Toward large-scale image segmentation on summit[C]. The 49th Intern-ational Conference on Parallel Processing-ICPP, Manh-attan: IEEE, 2020: 1-11.
[31]	Hohman F, Park H, Robinson C, et al. Summit: Scaling deep learning interpretability by visualizing activation and attribution summarizations[J]. IEEE transactions on visualization and computer graphics, 2019, 26(1): 1096-1106. doi: 10.1109/TVCG.2945
[32]	Zhang P, Yin D, Atkinson P M. Future Extreme Clim-ate Prediction in Western Jilin Province Based on Statis-tical DownScaling Model[C]. IGARSS 2019-2019 IEEE International Geoscience and Remote Sensing Symp-osium, Manhattan: IEEE, 2019: 9886-9889.
[33]	Torbicki M. Longtime prediction of climate-weather change influence on critical infrastructure safety and resilience[C]. The 2018 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), Manhattan: IEEE, 2018: 996-1000.
[34]	Shuman C A, Steffen K, Box J E, et al. A dozen years of temperature observations at the Summit: Central Green-land automatic weather stations 1987-99[J]. Journal of Applied Meteorology, 2001, 40(4): 741-752. doi: 10.1175/1520-0450(2001)040＜0741:ADYOTO＞2.0.CO;2
[35]	Anh Khoa T, Quang Minh N, Hai Son H, et al. Wireless sensor networks and machine learning meet climate cha-nge prediction[J]. International Journal of Communica-tion Systems, 2021, 34(3): e4687.
[36]	Guo S, Qiao W, Chen B, et al. Prediction and Abnor-mality Analysis of Climate Change Based on PCA-ARMA and PCC[C]. The 2020 IEEE International Conference on Networking, Sensing and Control (ICNSC), Manhattan: IEEE, 2020: 1-6.
[37]	Kappe C P, Böttinger M, Leitte H. Analysis of Decadal Climate Predictions with User‐guided Hierarchical Ens-emble Clustering[J]. Computer Graphics Forum, 2019, 38(3): 505-515. doi: 10.1111/cgf.13706
[38]	Belair S, Carrera M L, Abrahamowicz M, et al. Spaceb-orne L-Band Radiometry in Environment and Climate Change Canada (ECCC)’S Numerical Analysis and Prediction Systems[C]. IGARSS 2019-2019 IEEE Inter-national Geoscience and Remote Sensing Sympo-sium, Manhattan: IEEE, 2019: 7526-7528.
[39]	Massonnet F, Bellprat O, Guemas V, et al. Using climate models to estimate the quality of global observational data sets[J]. Science, 2016, 354(6311): 452-455. pmid: 27789838
[40]	Kurth T, Treichler S, Romero J, et al. Exascale deep learning for climate analytics[C]. The SC18:International Conference for High Performance Computing, Networ-king, Storage and Analysis, Manhattan: IEEE, 2018: 649-660.
[41]	Allen R M. Transforming earthquake detection?[J]. Sc-ience, 2012, 335(6066): 297-298.
[42]	Minson S E, Meier M A, Baltay A S, et al. The limits of earthquake early warning: Timeliness of ground motion estimates[J]. Science advances, 2018, 4(3): eaaq0504. doi: 10.1126/sciadv.aaq0504
[43]	Mapar J, Holtermann K, Legary J, et al. The role of inte-grated modeling and simulation in disaster preparedness and emergency preparedness and response: the SUMMIT platform[C]. The 2012 IEEE Conference on Technologies for Homeland Security (HST), Manhattan: IEEE, 2012: 117-122.
[44]	Luo H, Paal S G. Advancing post-earthquake structural evaluations via sequential regression-based predictive mean matching for enhanced forecasting in the context of missing data[J]. Advanced Engineering Informatics, 2021, 47(4):101202. doi: 10.1016/j.aei.2020.101202
[45]	Yu Z, Zhu K, Hattori K, et al. Borehole Strain Observ-ations Based on a State-Space Model and ApNe Analysis Associated With the 2013 Lushan Earthquake[J]. IEEE Access, 2021, 9: 12167-12179. doi: 10.1109/Access.6287639
[46]	Amin M S, Ahn H. Earthquake disaster avoidance lear-ning system using deep learning[J]. Cognitive Systems Research, 2021, 66: 221-235. doi: 10.1016/j.cogsys.2020.11.002
[47]	Hu Y, Yang H, Luan Z, et al. Massively scaling seismic processing on sunway taihulight supercomputer[J]. IEEE Transactions on Parallel and Distributed Systems, 2019, 31(5): 1194-1208. doi: 10.1109/TPDS.71
[48]	Ichimura T, Fujita K, Yamaguchi T, et al. A fast scalable implicit solver for nonlinear time-evolution earthquake city problem on low-ordered unstructured finite eleme-nts with artificial intelligence and transprecision compu-ting[C]. SC18:International Conference for High Perfo-rmance Computing, Networking, Storage and Anal-ysis, Manhattan: IEEE, 2018: 627-637.
[49]	Mazzucco W, Pastorino R, Lagerberg T, et al. Current state of genomic policies in healthcare among EU me-mber states: results of a survey of chief medical officers[J]. The European Journal of Public Health, 2017, 27(5): 931-937.
[50]	Chien S, Bashir R, Nerem R M, et al. Engineering as a new frontier for translational medicine[J]. Science trans-lational medicine, 2015, 7(281): 281 fs13-281fs13.
[51]	Kumar S, Huang C, Zheng G, et al. Scalable molecular dynamics with NAMD on the IBM Blue Gene/L system[J]. IBM Journal of Research and Development, 2008, 52(1):177-188. doi: 10.1147/rd.521.0177
[52]	Lv G F, Li M F, An H, et al. Distributed deep learning system for cancerous region detection on Sunway Tai-huLight[J]. CCF Transactions on High Performance Co-mputing, 2020, 2(4):1-14.
[53]	Sejdic E, Malandraki G A, Coyle J L. Computational deg-lutition: Using signal-and image-processing methods to understand swallowing and associated disorders[J]. IEEE signal processing magazine, 2018, 36(1): 138-146.
[54]	Joubert W, Weighill D, Kainer D, et al. Attacking the opioid epidemic: determining the epistatic and pleiot-ropic genetic architectures for chronic pain and opioid addiction[C]. SC18:International Conference for High Performance Computing, Networking, Storage and Anal-ysis, Manhattan: IEEE, 2018: 717-730.
[55]	Hush M R. Machine learning for quantum physics[J]. Science, 2017, 355(6325): 580. doi: 10.1126/science.aam6564 pmid: 28183936
[56]	Musser G. One of quantum physics’ greatest paradoxes may have lost its leading explanation[J/OL]. Science, 2020-09-12[2022-05-07]. https://www.science.org/cont-ent/article/one-quantum-physics-greatest-paradoxes-may-have-lost-its-leading-explanation.
[57]	Gao X, Zhang Z Y, Duan L M. A quantum machine lear-ning algorithm based on generative models[J]. Science advances, 2018, 4(12): eaat9004. doi: 10.1126/sciadv.aat9004
[58]	Gao P, Duan X, Zhang T, et al. Millimeter-scale and billion-atom reactive force field simulation on Sunway Taihulight[J]. IEEE Transactions on Parallel and Dis-tributed Systems, 2020, 31(12): 2954-2967.
[59]	Li K, Shang H, Zhang Y, et al. OpenKMC: a KMC desi-gn for hundred-billion-atom simulation using millions of cores on Sunway Taihulight[C]. Proceedings of the International Conference for High Performance Comp-uting, Networking, Storage and Analysis, Manh-attan: IEEE, 2019: 1-16.
[60]	Liu Z, Chu X S, Lv X, et al. Sunwaylb: Enabling extre-me-scale lattice boltzmann method based computing fluid dynamics simulations on sunway taihulight[C]. The 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS), Manhattan: IEEE, 2019: 557-566.
[61]	Zhang J, Zhou C, Wang Y, et al. Extreme-scale phase field simulations of coarsening dynamics on the sunway taihulight supercomputer[C]. SC’16:Proceedings of the International Conference for High Performance Comp-uting, Networking, Storage and Analysis, Manhattan: IEEE, 2016: 34-45.
[62]	Duan X, Gao P, Zhang T, et al. Redesigning LAMMPS for peta-scale and hundred-billion-atom simulation on Sunway TaihuLight[C]. SC18:International conference for high performance computing, networking, storage and analysis, Manhattan: IEEE, 2018: 148-159.
[63]	C Chang C C, Nicholson A N, Rinaldi E, et al. A percent-level determination of the nucleon axial coupling from quantum chromodynamics[J]. Nature, 2018, 558(7708): 91-94. doi: 10.1038/s41586-018-0161-8
[64]	Anbuvizhi R, Balakumar V. Credit/Debit Card Transa-ction Survey Using Map Reduce in HDFS and Imple-menting Syferlock to Prevent Fraudulent[J]. International Journal of Computer Science and Network Security (IJC-SNS), 2016, 16(11): 106.
[65]	Dai Y, Yan J, Tang X, et al. Online credit card fraud detection: a hybrid framework with big data technolo-gies[C]. The 2016 IEEE Trustcom/ BigDataSE / ISPA, Manhattan: IEEE, 2016: 1644-1651.
[66]	Li Z, Liu G, Jiang C. Deep representation learning with full center loss for credit card fraud detection[J]. IEEE Transactions on Computational Social Systems, 2020, 7(2): 569-579. doi: 10.1109/TCSS.6570650
[67]	Koppers L, Wormer H, Ickstadt K. Towards a systematic screening tool for quality assurance and semiautomatic fraud detection for images in the life sciences[J]. Science and engineering ethics, 2017, 23(4): 1113-1128. doi: 10.1007/s11948-016-9841-7 pmid: 27848190
[68]	Hong H J, Chuang J C, Hsu C H. Animation rendering on multimedia fog computing platforms[C]. The 2016 IEEE International Conference on Cloud Computing Technology and Science (CloudCom), Manhattan: IEEE, 2016: 336-343.
[69]	Zhang Y, Zhu Z, Cui H, et al. Small files storing and computing optimization in Hadoop parallel rendering[J]. Concurrency and Computation: Practice and Experience, 2017, 29(20): e3847. doi: 10.1002/cpe.v29.20
[70]	Wang A, Zhang A, Chan E, et al. A Review of Human Mobility Research Based on Big Data and Its Implication for Smart City Development[J]. International Journal of Geo-Information, 2020, 10(1):13.
[71]	de Assuncao M D, da Silva Veith A, Buyya R. Distributed data stream processing and edge computing: A survey on resource elasticity and future directions[J]. Journal of Network and Computer Applications, 2018, 103: 1-17. doi: 10.1016/j.jnca.2017.12.001
[72]	Mehmood Y, Ahmad F, Yaqoob I, et al. Internet-of-things-based smart cities: Recent advances and challenges[J]. IEEE Communications Magazine, 2017, 55(9): 16-24..
[73]	Bo T, Zhen C, Hefferman G, et al. Incorporating Intel-ligence in Fog Computing for Big Data Analysis in Smart Cities[J]. IEEE Transactions on Industrial Informatics, 2017, 13(5):2140-2150. doi: 10.1109/TII.2017.2679740
[74]	Jun S P, Yoo H S, Choi S. Ten years of research change using Google Trends: From the perspective of big data utilizations and applications[J]. Technological Forecas-ting and Social Change, 2018, 130(5):69-87.
[75]	Sapountzi A, Psannis K E. Social networking data analysis tools & challenges[J]. Future Generation Computer Sys-tems, 2018, 86: 893-913.
[76]	Stavrinides G L, Karatza H D. The impact of data locality on the performance of a SaaS cloud with real-time data-intensive applications[C]. The 2017 IEEE/ACM 21st International Symposium on Distributed Simulation and Real Time Applications (DS-RT), Manhattan: IEEE, 2017: 1-8.
[77]	Abramson D, Jin C, Luong J, et al. A BeeGFS-based caching file system for data-intensive parallel computing[C]. Asian Conference on Supercomputing Frontiers, Cham: Springer, 2020: 3-22.
[78]	Zhang X, Wang Y, Wang Q, et al. A new approach to double i/o performance for ceph distributed file system in cloud computing[C]. The 2019 2nd International Con-ference on Data Intelligence and Security (ICDIS), Man-hattan: IEEE, 2019: 68-75.
[79]	Iannone F, Ambrosino F, Bracco G, et al. CRESCO EN-EA HPC clusters: a working example of a multifabric GPFS Spectrum Scale layout[C]. The 2019 International Conference on High Performance Computing & Simul-ation (HPCS), Manhattan: IEEE, 2019: 1051-1052.
[80]	Braam P. The Lustre storage architecture[J/OL]. arXiv preprint arXiv:1903.01955, 2019-03-05 [2022-05-07]. https://arxiv.org/abs/1903.01955.
[81]	Awad M, Menascé D A. iModel: Automatic Derivation of Analytic Performance Models[J]. ACM Transactions on Modeling and Performance Evaluation of Computing Systems (TOMPECS), 2020, 5(2): 1-30.
[82]	Williams S, Waterman A, Patterson D. Roofline: an insi-ghtful visual performance model for multicore archit-ectures[J]. Communications of the ACM, 2009, 52(4): 65-76. doi: 10.1145/1498765.1498785
[83]	Tsafrir D, Etsion Y, Feitelson D G. Backfilling Using System-Generated Predictions Rather than User Runt-ime Estimates[J]. IEEE Transactions on Parallel & Distri-buted Systems, 2007, 18(6):789-803.
[84]	Gibbons R. A historical application profiler for use by parallel schedulers[C]. Workshop on Job Scheduling Stra-tegies for Parallel Processing, Berlin:Springer, 1997: 58-77.
[85]	Smith W, Foster I, Taylor V. Predicting application run times using historical information[C]. Workshop on Job Scheduling Strategies for Parallel Processing, Berlin:Springer, 1998: 122-142.
[86]	Kapadia N H, Fortes J A B, Brodley C E. Predictive app-lication-performance modeling in a computational grid environment[C]. Proceedings. The Eighth International Symposium on High Performance Distributed Computing (Cat. No. 99TH8469), Manhattan: IEEE, 1999: 47-54.

名称	文件系统	存储访问接口	存储容量	算力类型	存储带宽
富岳	Lustre+BeeGFS	POSIX	1EB	CPU	10TB/s
Summit	IBM Spectrum Scale	POSIX SWIFT HDFS	250PB	CPU GPU	2.5TB/s
神威太湖之光	Lustre+SWGFS	POSIX	10PB	CPU	341 GB/s
天河2A	Lustre+H2FS	POSIX HDFS	19PB	CPU MIC GPU	1TB/s

应用名称	存储规模	IOPS要求	AI库	大数据接口	带宽要求	可靠性要求	安全要求	实时性要求
E级深度学习气候分析	3.5TB	高	需要	不需要	高	高	中	高
演化地震城市问题求解器	TB-PB级	高	需要	不需要	高	高	中	低
基因组学分析的药物成瘾症	10TB	高	需要	需要	高	高	高	低
量子动力学模拟	TB级	高	需要	不需要	高	高	中	低

应用名称	存储规模	IOPS要求	带宽要求	AI库	大数据接口	可靠性要求	安全要求	实时性要求
信用卡欺诈检测	PB至EB级别	高	高	需要	需要	高	高	高
动漫渲染	GB至TB级别	高	高	不需要	不需要	高	中	低
智慧城市	PB级	高	高	需要	需要	高	高	低
知识获取	ZB级	高	高	需要	需要	高	中	低

硬件	硬件型号	功耗（W）	计算能力（FLOPS）	开源软件支持	并行算法开发难度	开发周期
CPU	Ryzen ThreadRipper_3970X	280	4062G SP	好	易	短
GPU	TeslaV100	235	15T SP	好	中等	短
MIC	KNL	225-300	3T DP	少	易	中
FPGA	UltraScale+	27	4903G SP	少	难	较长
ASIC	VPU	1	150G SP	少	难	长
TPU	TPU v2	118	180T SP	好	难	较长

名称	类型
Gray	数据灰度运算
Cube	数据立方运算
Globalreduce	全局reduce运算
Cputest	选择排序
SumTest	矩阵相加
Transpose	矩阵转置
Histunique	独立直方运算