Search academic texts

Information × Registration Number 0225U004782, (0122U201569) , R & D reports Title High-performance multiprocessor systems: features of design, study of efficiency estimates, application for solving applied problems. popup.stage_title Високопродуктивні багатопроцесорні системи: особливості конструювання, дослідження оцінок ефективності, застосування до розв’язування прикладних задач. Head Shvachych Hennadii H., Доктор технічних наук Registration Date 18-12-2025 Organization National Technical University Dnipro Polytechnic popup.description1 Improving the architecture of modular multiprocessor computing systems and increasing their efficiency by developing parallel computing complexes adapted for the class of problems under study, as well as introducing new numerical-analytical algorithms and methods based on them, which improve the accuracy, performance and reliability of data processing. popup.description2 Improvements to the architecture of modular multiprocessor computing systems are proposed, along with enhancements to their efficiency through the development of parallel computing complexes adapted to the class of problems under study, as well as the implementation of new numerical-analytical algorithms and methods. These advancements increase the accuracy, performance, and reliability of data processing. To create a technology for parallelizing algorithms for solving applied problems, the basic principles of linear algebra, graph theory, the theory of ordinary differential equations and partial differential equations, the theory of difference schemes, and key methods of mathematical analysis (differentiation of functions, Taylor series expansions, etc.) were applied. The developed multiprocessor computing technologies are aimed at solving a wide range of applied problems and make it possible to improve the accuracy, performance, and efficiency of mathematical modeling. The obtained results make it possible to: use the model of the developed multiprocessor computing system as an integrated environment that supports distributed design of software and hardware for mathematical modeling of a specific class of applied problems; accelerate the process of modifying a distributed computing system to ensure the required performance for solving a given class of applied problems through the appropriate selection of standard modules; and speed up the recovery of a personal multiprocessor computing system by automatically replacing failed modules with functional ones within the proposed system configuration constructed from standard modules Product Description popup.authors Borys I. Moroz Shyrin Artem L. Iryna M. Udovyk Leonid V. Kabak Oleh V. Ivanchenko Ivan S. Laktionov Mykhailo O. Aleksieiev Leonid I. Meshcheriakov Andrii A. Martynenko Dmytro M. Moroz Olha S. Shevtsova Alona T. Khar Pavlo O. Ishchuk Kateryna S. Rodna popup.nrat_date 2025-12-18 Close