علوم رایانش و فناوری اطلاعات

مروری بر معیارهای محک آزمون و ارزیابی سامانه‌‏های پردازشی فوق‏‌سریع به همراه نصب، پیکربندی و تحلیل معیارهای منتخب و مطالعه موردی در سامانه پردازشی فوق‌‏سریع نمونه

نوع مقاله : مقاله پژوهشی

نویسندگان

پژوهشگاه ارتباطات و فناوری اطلاعات، تهران، ایران

چکیده

در این مقاله به بررسی معیارهای محک رایج برای آزمون و ارزیابی سامانه‌های پردازشی فوق‌سریع پرداخته شده است. بدین منظور، با مرور معیارهای مختلف رتبه‌بندی‌ سامانه‌های پردازشی فوق‌سریع، روش‌های آزمون و ارزیابی معیارهای منتخب برای منابع پردازشی، حافظه، شبکه و دیسک معرفی و نحوه راه‌اندازی، پیکربندی اولیه و اجرای هر یک شرح داده می‌شود. همچنین با راه‌اندازی یک سامانه آزمایشی، خروجی هر یک از معیارها به تفکیک ارائه و تحلیل می‌شود. در انتها، به منظور نشان دادن کاربردپذیری و داشتن درک صحیحی از مقدار خروجی معیارها در عمل، تحلیل نتایج حاصل از معیارها برای یک سامانه پردازشی فوق‌سریع واقعی ارائه می‌گردد. از آنجا که معیارهای در نظر گرفته شده در این مقاله از معتبرین معیارهای ارزیابی بوده و نیز دستورات آماده‌سازی محیط و نصب و اجرای معیارها با جزئیات کامل ارائه شده است، از محتویات این مقاله می‌توان برای انتخاب معیار، و نصب و راه‌اندازی آن برای آزمون و ارزیابی سامانه‌های پردازشی فوق‌سریع در ابعاد واقعی و به صورت عملیاتی استفاده نمود. 

کلیدواژه‌ها

  • [1] J. J. Dongarra, P. Luszczek, and A. Petitet, “The LINPACK Benchmark: past, present and future,” Concurrency and Computation: Practice and Experience, vol. 15, no. 9, pp. 803–820, Aug. 2003, doi: 10.1002/cpe.728.
  • [2] “HPL - A Portable Implementation of the High-Performance Linpack Benchmark for Distributed-Memory Computers.” https://www.netlib.org/benchmark/hpl/ (accessed Sep. 26, 2021).
  • [3] P. Gepner, V. Gamayunov, and D. L. Fraser, “Effective Implementation of DGEMM on Modern Multicore CPU,” Procedia Computer Science, vol. 9, pp. 126–135, 2012, doi: 10.1016/j.procs.2012.04.014.
  • [4] “NAS Parallel Benchmarks.” https://www.nas.nasa.gov/software/npb.html (accessed Sep. 26, 2021).
  • [5] D. H. Bailey et al., “The NAS parallel benchmarks---summary and preliminary results,” in Proceedings of the 1991 ACM/IEEE conference on Supercomputing - Supercomputing ’91, 1991, pp. 158–165. doi: 10.1145/125826.125925.
  • [6] ASC Community, “Supercomputer System Performance Evaluation Methods,” in The Student Supercomputer Challenge Guide, Singapore: Springer Singapore, 2018, pp. 81–115. doi: 10.1007/978-981-10-3731-3_5.
  • [7] J. Ang, B. Barrett, K. Wheeler, and R. Murphy, “Introducing the graph 500,” Sep. 2010.
  • [8] “Graph 500 | large-scale benchmarks.” https://graph500.org/ (accessed Sep. 26, 2021).
  • [9] “HPL-AI Mixed-Precision Benchmark.” https://hpl-ai.org/ (accessed Sep. 26, 2021).
  • [10] “MEMORY BANDWIDTH: STREAM BENCHMARK PERFORMANCE RESULTS.” https://www.cs.virginia.edu/stream/ (accessed Sep. 26, 2021).
  • [11] V. Aggarwal, Y. Sabharwal, R. Garg, and P. Heidelberger, “HPCC RandomAccess benchmark for next generation supercomputers,” in 2009 IEEE International Symposium on Parallel & Distributed Processing, May 2009, pp. 1–11. doi: 10.1109/IPDPS.2009.5161019.
  • [12] R. Rabenseifner and A. Koniges, “The Effective I/O Bandwidth Benchmark (b_eff_io),” Sep. 2000.
  • [13] P. Luszczek et al., “Introduction to the HPC Challenge Benchmark Suite,” Sep. 2004.
  • [14] S. Potluri, “The MVAPICH Project: Evolution and Sustainability of an Open Source Production Quality MPI Library for HPC.” figshare, Sep. 2013. doi: 10.6084/m9.figshare.790743.v3.
  • [15] “IOR and mdtest.” https://github.com/hpc/ior (accessed Sep. 26, 2021).
  • [16] “MDTest - Lustre Wiki.” https://wiki.lustre.org/MDTest (accessed Sep. 26, 2021).
  • [17] “The Linpack Benchmark | TOP500.” https://www.top500.org/project/linpack/ (accessed Sep. 26, 2021).
  • [18] J. J. Dongarra, “The LINPACK Benchmark: An explanation,” Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 297 LNCS, pp. 456–474, 1987, doi: 10.1007/3-540-18991-2_27.
  • [19] LawsonC. L., HansonR. J., KincaidD. R., and KroghF. T., “Basic Linear Algebra Subprograms for Fortran Usage,” ACM Transactions on Mathematical Software (TOMS), vol. 5, no. 3, pp. 308–323, Sep. 1979, doi: 10.1145/355841.355847.
  • [20] R. F. Barrett, T. H. F. Chan, E. F. D’Azevedo, E. F. Jaeger, K. Wong, and R. Y. Wong, “Complex version of high performance computing LINPACK benchmark (HPL),” Concurrency and Computation: Practice and Experience, vol. 22, no. 5, pp. 573–587, Apr. 2010, doi: 10.1002/CPE.1476.
  • [21] P. Luszczek and J. Dongarra, “Analysis of various scalar, vector, and parallel implementations of RandomAccess *,” 2010.
  • [22] J. Dongarra, M. A. Heroux, and P. Luszczek, “A new metric for ranking high-performance computing systems,” National Science Review, vol. 3, no. 1, pp. 30–35, Mar. 2016, doi: 10.1093/NSR/NWV084.
  • [23] J. Dongarra and P. Luszczek, “HPC challenge: Design, history, and implementation highlights,” Contemporary High Performance Computing: From Petascale toward Exascale, pp. 13–30, Jan. 2013, doi: 10.1201/9781351104005-2.
  • [24] H. W. Meuer, “The TOP500 Project. Looking Back over 15 Years of Supercomputing Experience,” vol. 31, no. 2, pp. 122–132, Jun. 2008, doi: 10.1515/PIKO.2008.0022.
  • [25] V. Turchenko, L. Grandinetti, G. Bosilca, and J. J. Dongarra, “Improvement of parallelization efficiency of batch pattern BP training algorithm using Open MPI,” Procedia Computer Science, vol. 1, no. 1, pp. 525–533, 2010, doi: 10.1016/J.PROCS.2010.04.056.
  • [26] D. Yokoyama, B. Schulze, F. Borges, and G. Mc Evoy, “The survey on ARM processors for HPC,” The Journal of Supercomputing 2019 75:10, vol. 75, no. 10, pp. 7003–7036, Jun. 2019, doi: 10.1007/S11227-019-02911-9.
  • [27] Z. Jiang et al., “HPC AI500: A Benchmark Suite for HPC AI Systems,” Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11459 LNCS, pp. 10–22, Dec. 2018, doi: 10.1007/978-3-030-32813-9_2.
  • [28] T. Muhammed, R. Mehmood, A. Albeshri, and F. Alsolami, “HPC-Smart Infrastructures: A Review and Outlook on Performance Analysis Methods and Tools,” EAI/Springer Innovations in Communication and Computing, pp. 427–451, 2020, doi: 10.1007/978-3-030-13705-2_18.
  • [29] A. Snavely, G. Chun, H. Casanova, R. F. van der Wijngaart, and M. A. Frumkin, “Benchmarks for grid computing,” ACM SIGMETRICS Performance Evaluation Review, vol. 30, no. 4, pp. 27–32, Mar. 2003, doi: 10.1145/773056.773062.
  • [30] E. Gabriel et al., “Open MPI: Goals, Concept, and Design of a Next Generation MPI Implementation,” Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 3241, pp. 97–104, 2004, doi: 10.1007/978-3-540-30218-6_19.
  • [31] L. Dagum and R. Menon, “OpenMP: an industry standard API for shared-memory programming,” IEEE Computational Science and Engineering, vol. 5, no. 1, pp. 46–55, 1998, doi: 10.1109/99.660313.
  • [32] “GEO ExPro - Supercomputers for Beginners - Part III. GPU-Accelerated Computing.” https://www.geoexpro.com/articles/2016/03/supercomputers-for-beginners-part-iii-gpu-accelerated-computing (accessed Sep. 27, 2021).
  • [33] “LINPACK benchmarks - Wikipedia.” https://en.wikipedia.org/wiki/LINPACK_benchmarks (accessed Sep. 27, 2021).
  • [34] “HPL Tuning.” https://www.netlib.org/benchmark/hpl/tuning.html (accessed Sep. 27, 2021).
  • [35] “How do I tune my HPL.dat file? - Advanced Clustering Technologies.” https://www.advancedclustering.com/act_kb/tune-hpl-dat-file/ (accessed Sep. 27, 2021).
  • [36] R. P. Brent, “The LINPACK benchmark on the Fujitsu FAP 1000,” in [Proceedings 1992] The Fourth Symposium on the Frontiers of Massively Parallel Computation, pp. 128–135. doi: 10.1109/FMPC.1992.234897.
  • [37] “Frequently Asked Questions | TOP500.” https://www.top500.org/resources/frequently-asked-questions/ (accessed Sep. 27, 2021).
  • [38] “RandomAccess Rules.” https://icl.utk.edu/projectsfiles/hpcc/RandomAccess/ (accessed Sep. 26, 2021).
  • [39] “IOR Introduction.” https://ior.readthedocs.io/en/latest/intro.html (accessed Sep. 26, 2021).
  • [40] “November 2021 | TOP500.” https://top500.org/lists/top500/2021/11/ (accessed Dec. 01, 2021).
  • [41] https://www.dell.com/support/kbdoc/de-de/000133009/hpc-synthetic-benchmark-performance-using-2nd-generation-intel-xeon-scalable-processors-stream-hpl-and-hpcg?lang=en
  • [42] V. Aggarwal, Y. Sabharwal, R. Garg and P. Heidelberger, "HPCC RandomAccess benchmark for next generation supercomputers," 2009 IEEE International Symposium on Parallel & Distributed Processing, 2009, pp. 1-11, doi: 10.1109/IPDPS.2009.5161019.
  • [43] Florida S. Uni.: https://acct.rcc.fsu.edu/doc/hpc-benchmarks
علوم رایانش و فناوری اطلاعات
دوره 20، شماره 2
پاییز و زمستان 1401
آذر 1401