Article

Article title HYBRID SUPERCOMPUTER PLATFORM AND APPLICATIONS PROGRAMMING FOR THE SOLUTION OF CONTINUOUS MECHANICS PROBLEMS BY GRID METHODS
Authors S.V. Polyakov, Yu.N. Karamzin, O.A. Kosolapov, T.A. Kudryashova, S.A. Soukov
Section SECTION III. USE OF THE SUPER COMPUTER IN MATHEMATICAL MODELLING
Month, Year 06, 2012 @en
Index UDC 519.688
DOI
Abstract The developing of parallel applications for the solution of continuum mechanics problems on modern computer systems with hybrid architecture (including the central and graphical processors) was considered. For the solving of the problem the conception of hybrid parallel computations was formulated. This conception includes the analysis of both architecture of modern hybrid computer and its software and the special ways for the construction of parallel programs. In particular, three main models of the parallel calculations, using graphic processors incidentally, constantly or in a close sheaf with the central processors were offered. Specific problems of realization of grid numerical algorithms on hybrid computers were also considered. The example of implementation of the finite-volume scheme on unstructured grid at the decision of systems of the equations of Euler and Navier-Stokes on graphic accelerators was given.

Download PDF

Keywords Сontinuum mechanics; mathematical modeling; numerical approach based on grid technique; parallel algorithms; software tools for hybrid computer systems.
References 1. Сандерс Дж., Кэндрот Э. Технология CUDA в примерах: введение в программирование графических процессоров. - М.: ДМК Пресс, 2011. - 232 с.
2. NVIDIA CUDA C Programming Guide. Version 4.0. Santa Clara (CA, USA): NVIDIA Corporation, 2011. - 187 p.
3. Боресков А.В., Харламов А.А. Основы работы с технологией CUDA. - М.: ДМК Пресс, 2010. - 232 с.
4. Logan S. Cross-platform development in C++ : building Mac OS X, Linux, and Windows applications. Crawfordsville (Indiana, USA): RR Donnelly, 2007. - 575 p.
5. Стефенс Д.Р., Диггинс К., Турканис Д., Когсуэлл Д. С++. Сборник рецептов. - М.: КУДИЦ-ПРЕСС, 2007. - 624 с.
6. Ортега Дж. Введение в параллельные и векторные методы решения линейных систем. - М.: Мир, 1971. - 367 с.
7. Воеводин В.В., Воеводин Вл.В. Параллельные вычисления. - СПб.: БХВ-Петербург, 2002. - 608 с.
8. Dongarra J., Foster I., Fox J. et al. Sourcebook of Parallel Computing. San Francisсo (CA, USA): Elsevier Science, 2003. - 852 p.
9. Гергель В.П. Теория и практика параллельных вычислений. СПб.: "Интернет-университет информационных технологий - ИНТУИТ.ру", "БИНОМ. Лаборатория знаний", 2007. - 424 c.

Comments are closed.