AGGREGATION OF COMPUTING CHANNELS BASED ON THE NVIDIA CUDA PLATFORM FOR CONTROL MODES OF COMPONENTS OF TECHNOLOGICAL SYSTEMS
DOI:
https://doi.org/10.32782/IT/2022-2-10Keywords:
multiprocessor systems, aggregation, speed, memory, computing nodes, technological process, processors, graphic objects, non-graphic computing, annealing, spheroidizationAbstract
Today, practice poses problems, the solution of which by well-known standard approaches quite often represents a significant problem, which can be solved only by using multiprocessor computer technologies. In turn, one of the fundamental features of the application of these technologies is reduced to increasing the productivity and speed of calculations. At the same time, the significant performance of calculations allows the solution of multidimensional problems, as well as problems that require a significant amount of processor time. Speed operation allows you to effectively manage not only technological processes, but also provides for the creation of prerequisites for the development of promising and innovative technological processes. Therefore, the application of high-performance computing is an urgent and priority problem today. The goal of the work is to improve the structure and increase the performance of a multiprocessor computer system by aggregating computing channels based on the NVIDIA CUDA platform for control modes of technological process components. The proposed approach made it possible not only to increase the efficiency of parallelization, but also to significantly reduce the calculation time. In the given development of a multiprocessor system, two NVIDIA GeForce GTX 1080 video cards were "connected". This approach is aimed not only at a significant increase in computing performance, but also at a significant decrease in latency and significant unloading of the system bus. Compared to the known approach, due to the application of the software-hardware architecture of parallel computing from the NVIDIA corporation based on the CUDA platform, it was possible to increase the volume of video memory by 16 GB on each computing node of the multiprocessor system, as well as to increase the overall performance of the system node by 350 GFL. The practical value of the conducted research is aimed at solving the problem of intensification of spheroidizing annealing of a long steel product. The direct technological process of heat treatment of metal acquires such advantages as high productivity, a significant reduction in energy consumption, and allows control of technological parameters by the length and cross-sectional area of the metal.
References
Швачич Г.Г., Коломоец Н.А., Соболенко А.В. Исследование и разработка теплового режима отжига бунтов в шахтной печи. Теплотехника и энергетика в металлургии. Дніпро. 2002. С. 14–18.
Швачич Г.Г., Соболенко А.В., Соболенко М.А. Исследование температурного режима работы печей отжига шахтного типа бунтов проволоки. Теория и практика металлургии. 2003. № 1. С. 59–62.
Shvachych G., Pobochij I., Sazonova M., Bilyi O., Moroz D. Intelligent decision support system. International Academy Journal Web of Scholar. 2021. № 2 (52). P. 1–9. 4.
Швачич Г.Г., Соболенко А.В., Протопопов Д.В., Чуев А.В. Информационная система сопровождения трубопрокатного агрегата 140 со станами типа тандем. Теория и практика металлургии. 2003. № 5–6. С. 76–82.
Швачич Г.Г., Шмукин А.А. Особенности конструирования параллельных вычислительных алгоритмов для ПЭВМ в задачах тепло- и массообмена. Восточно-Европейский журнал передовых технологий. 2004. № 3. С. 42–47.
Долженков И.Е. Термическая и деформационно – термическая обработка металлопроката. Теория и практика металлургии. 2002. № 3. – С.30–36.
Іващенко В.П., Швачич Г.Г., Соболенко М.О. Новітні металургійні технології на основі використання високопродуктивних багатопроцесорних обчислювальних комплексів. Теория и практика металлургии. 2013. № 3–4 . С.78–87.
Долженков И.Е. Влияние пластической деформации и других предобработок на сфероидизацию карбидов в сталях. Теория и практика металлургии. 2007. № 1. С. 66–68.
Соболенко М.А. О проблеме математического моделирования параметров скоростной сфероидизирующей обработки углеродистых и низколегированных сталей. Сучасні проблеми металургії. 2007. № 10. С. 123–137.
Гуль Ю.П., Соболенко М.А. Интенсификация процесса сфероидизирующего отжига стали в поточних линях. Строительство, материалловедение, машиностроение. 2011. № 58. С. 197–202.
Патент №143971 Україна МПК C21D 1/26 (2006.01) Установка для інтенсифікації сфероїдизуючого відпалювання сталі / Іващенко В. П., Швачич Г. Г., Соболенко М.О., Гуль Ю.П., Соболенко О.В., Кокашинська Г.В., Мороз Д. М. №u202000940; заявл. 14.02.2020; опубл. 25.08.2020, Бюл. № 16.
Shvachych G., Moroz B., Martynenko A., Hulina I., Busygin V., Moroz D. Model of Speed Spheroidization of Metals and Alloys Based on Multiprocessor Computing Complexes. Machine Learning for Predictive Analysis. Networks and Systems. Springer. 2020. P. 33–41.
Shvachych G., Pobochii I., Khokhlova T., Kholod A., Moroz D. Multiprocessor Computing based Parallel Structures of Mathematical Models of Tridiagonal Systems. 5th International Conference on Inventive Computation Technologies. 2020. P. 1031–1035.
Moroz D. Numerical-analytical method for distributed modelling of applied tasks. Science and education : Proceedings of VII International Scientific and Practical Conference. Kyoto, Japan. 2021. P. 128–136.
Мороз Д. М. Моделирование максимально параллельных структур алгоритмов решения тепловых задач. Сучасні проблеми металургії. Дніпро. 2021. № 24. С. 98–109.
