DISCRETE-EVENT METHOD COMPUTATIONS ORGANIZING FOR PROCESSING LARGE SPARSE UNSTRUCTURED MATRIXES ON RCS
Abstract
Increasing models complexity objects and processes study, in different sphere of science and technology, set up plenty issues to necessary to use high-performance computing systems. Arrays matrix processing by cluster multiprocessor computing systems in conjunction special methods aimed at organizing parallel computations, basically obtain computing performance system is quite high. However, that computational efficiency is not observed for all types of matrices. Matrix structure be in a position contain large amount of insignificant elements, large dimension and unstructured portrait. Calculation execute for described kind of matrices on cluster multiprocessor computing system couldn't achieve close peak performance. Considering that processing methods leave out the complex structure of the matrix being processed. As a result, the performance of the system is significantly reduced. The development of cluster MCS methods doesn't allow for full ensure high performance for class of problems processing of large sparse unstructured matrices. Rigid architecture of processor commutation net doesn’t take into account the peculiarities of such matrices, and lead to non-uniformity loading processor. To achieve performance close the peak for tasks large sparse unstructured matrices processing necessary to use reconfigurable compu-ting systems. RCS architecture allows adapting computation structure to the problem solved. This makes it possible to organize pipeline processing, such a way that computational resource RCS used only for informational significant operations. In addition using generally accepted methods for structural organization of high-performance computing for RCS, it is necessary to develop a format for storing and transferring large sparse unstructured matrices, to determine the principles of constructing basic matrix macro-operations and the possibility of organizing composite dis-crete-event matrix functions for solving applied problems. Сconsequently method founding laid allows organizing computations operands, which are large sparse unstructured matrices. The application this method for organizing computations can significantly increase productivity, and provide an increase in the efficiency of such a system
References
1. Tikhonov A.N., Samarskiy A.A. Uravneniya matematicheskoy fiziki [Equations of mathemati-cal physics]. Moscow: Izd-vo Moskovskogo universiteta, 1999. 6 ed., 798 p. Nauchnaya biblioteka dissertatsiy i avtoreferatov disserCat [Dissercat Scientific Library of Dissertations and Abstracts]. Available at: http://www.dissercat.com/content/razrabotka-i-issledovanie-ekonomichnykh-algoritmov-resheniya-setochnykh-zadach-naklastere-r#ixzz5WvO2qG2e (ac-cessed 23 November 2018).
2. Kolodziej Scott P., Mohsen Aznaveh, Matthew Bullock, Jarrett David, Davis Timothy A., Mat-thew Henderson, Yifan Hu, and Read Sandstrom. The SuiteSparse Matrix Collection Website Interface, Journal of Open Source Software 4, 35 (March 2019), pp. 1244-1248. DOI: https://doi.org/10.21105/joss.01244 (accessed 02 December 2021).
3. Параллельные вычисления CUDA, NVIDIA Corporation, 2018. Available at: http://www.nvidia.ru/object/cuda-parallel-computing-ru.html (accessed 14 December 2021).
4. Chungz E.S., Davisz J.D., Kestury S. An FPGA Drop-In Replacement for Universal Matrix-Vector Multiplication. Portland: Workshop on the Intersections of Computer Architecture and Reconfigurable Logic, 2012, pp. 1-6.
5. Yang C., Buluc A., Owens J. Design Principles for Sparse Matrix Multiplication on the GPU, Turin: International European Conference on Parallel and Distributed Computing, 2018, pp. 12.
6. Georgopoulos L., Sobczyk A., Christofidellis D., Dolfi M., Auer C., Staar P., Bekas C. Enhancing multi-threaded sparse matrix multiplication for knowledge graph oriented algorithms and analyt-ics IBM Research. Zurich Säumerstrasse 4 CH-8803 Rüschlikon Switzerland, 2019, 11 p.
7. Podoprigora A.V. Metody i sredstva obrabotki bol'shikh razrezhennykh matrits na rekonfiguriruemoy vychislitel'noy sisteme [Methods and means of processing large sparse ma-trices on a reconfigurable computing system], XVI Ezhegodnaya molodezhnaya nauchnaya konferentsiya «YUg Rossii: vyzovy vremeni, otkrytiya, perspektivy»: Mater. konferentsii (g. Rostov-na-Donu, 13–28 aprelya 2020 g.) [XVI Annual Youth Scientific Conference "South of Russia: challenges of the time, discoveries, prospects": materials of the conference (Rostov-on-Don, April 13-28, 2020)]. Rostov-on-Don: Izd-vo YuNTS RAN, 2020, 168 p. ISBN 978-5-4358-0196-5.
8. Kunchum R. On Improving Sparse Matrix-Matrix Multiplication on GPUs (Thesis). The Ohio State University, 2017, pp. 36-42 Available at: https://etd.ohiolink.edu/!etd.send_file? acces-sion=osu1492694387445938&disposition=inline.
9. Sorokin D.A. Metody resheniya zadach s peremennoy intensivnost'yu potokov dannykh na rekonfiguriruemykh vychislitel'nykh sistemakh: disc. … kand. tekhn. nauk [Methods for solv-ing problems with variable intensity of data flows on reconfigurable computing systems: cand. of eng. sc. diss.]: 05.13.11 protected: 15.06.12: approved: 11.03.13. Taganrog, 2013, 168 p. 005043774.
10. Kalyaev I.A., Levin I.I., Semernikov E.A., Shmoylov V.I. Rekonfiguriruemye mul'tikonveyernye vychislitel'nye struktury [Reconfigurable multiconveyor computing structures], under the gen-eral ed. by I.A. Kalyaeva. 2nd ed. revised and expanded. Rostov-on-Don: Izd-vo YuNTS RAN, 2009б 344 p.
11. Podoprigora A.V., Chekina M.D. Reshenie razrezhennykh SLAU bol'shoy i sverkhbol'shoy razmernosti mnogosetochnym metodom na RVS [The solution of sparse SLOWS of large and extra-large dimensions by the multigrid method on RVS], Izvestiya YuFU. Tekhnicheskie nauki [Izvestiya SFedU. Engineering Sciences], 2018, No. 8 (202), pp. 212-221.
12. Dordopulo A.I. Kalyaev I.A., Levin I.I., Semernikov E.A. Semeystvo mnogoprotsessornykh vychislitel'nykh sistem s dinamicheski perestraivaemoy arkhitekturoy [A family of multipro-cessor computing systems with dynamically tunable architecture], Mnogoprotsessornye vychislitel'nye i upravlyayushchie sistemy: Mater. nauchno-tekhnicheskoy konferentsii [Multi-processor computing and control systems: Proceedings of the scientific and technical confer-ence]. Taganrog, 2007, pp. 11-17.
13. Kalyaev I.A., Levin I.I., Semernikov E.A., Dordopulo A.I. Rekonfiguriruemye vychislitel'nye sistemy na osnove PLIS semeystva VIRTEX-6 [Reconfigurable computing systems based on FPGAs of the VIRTEX-6 family], Parallel'nye vychislitel'nye tekhnologii (PAVT’2011): Tr. mezhdunarodnoy nauchnoy konferentsii [Parallel computing technologies (PAVT'2011): Proceedings of the International Scientific Conference], 2011, pp. 203-211.
14. Guzik V.F., Kalyaev I.A., Levin I.I. Rekonfiguriruemye vychislitel'nye sistemy [Reconfigurable computing systems], ed. by I.A. Kalyaeva. Taganrog: Izd-vo YuFU, 2016, 472 p.
15. Kalyaev A.V., Levin I.I. Modul'no-narashchivaemye mnogoprotsessornye sistemy so strukturno-protsedurnoy organizatsiey vychisleniy [Modular-stackable multiprocessor systems with structural and procedural organization of computing]. Moscow: Yanus-K, 2003, 380 p.
16. Pelipets A.V. Metody i sredstva resheniya zadach lineynoy algebry na vysokoproizvoditel'nykh rekonfiguriruemykh vychislitel'nykh sistemakh: diss. … kand. tekhn. nauk [Methods and means of solving linear algebra problems on high-performance reconfigurable computing sys-tems: cand. of eng. sc. diss.]. Taganrog, 2016, 199 p.
17. Podoprigora A.V. Upravlenie protsessom obrabotki razrezhennykh matrits v diskretno-sobytiynykh matrichnykh operatsiyakh [Managing the process of processing sparse matrices in discrete-event matrix operations], XIV Vserossiyskaya mul'tikonferentsiya po problemam upravleniya (MKPU-2021): Mater. XIV mul'tikonferentsii (Divnomorskoe, Gelendzhik, 27 sentyabrya – 2 oktyabrya 2021 g.) [XIV All-Russian Multi-conference on Management Prob-lems (MCPU-2021): Materials of the XIV multi-conference (Divnomorskoye, Gelendzhik, September 27 - October 2, 2021)]: In 4 vol. Vol. 2, Editorial board: I.A. Kalyaev, V.G. Peshekhonov i dr. Rostov-on-Don; Taganrog: Izd-vo YuFU, 2021, pp. 276-278. ISBN 978-5-9275-3846-1.
18. Kleynrok L. Teoriya massovogo obsluzhivaniya [Theory of queuing]. Moscow: Mashinostroenie, 1979, 432 p.
19. Podoprigora A.V., Sorokin D.A. Upravlenie potokami dannykh pri reshenii bol'shoy razrezhennoy slau mnogosetochnym metodom na rekonfiguriruemykh vychislitel'nykh sistemakh [Managing data flows when solving a large sparse SLA by a multigrid method on reconfigurable computing systems], XII mul'tikonferentsiya po problemam upravleniya (MKPU-2019): Mater. XII mul'tikonferentsii (Divnomorskoe, Gelendzhik, 23–28 sentyabrya 2019 g.) [XII Multi-conference on Management Problems (MCPU-2019): Materials of the XII multi-conference (Divnomorskoye, Gelendzhik, September 23-28, 2019)]: In 4 vol. Vol. 3, Ed-itorial board: I.A. Kalyaev, V.G. Peshekhonov i dr. Rostov-on-Don, Taganrog: Izd-vo YuFU, 2019, pp. 120-122. ISBN 978-5-9275-3188-2.
20. Konnov A.L., Ushakov Yu.A. Metody rascheta pokazateley proizvoditel'nosti setey EVM s neodnorodnym trafikom [Methods for calculating performance indicators of computer net-works with heterogeneous traffic]. Orenburg: OGU, 2013, pp. 10-16.
