To obtain access to full text of journal and articles you must register!
- Article name
- SOFTWARE AND HARDWARE PARALLELIZATION FOR COMPUTATIONAL FLUID DYNAMICS PROBLEMS
- Authors
- Zhilenkov A. A., , zhilenkovanton@gmail.com, St. Petersburg State Marine Technical University, St. Petersburg, Russia
Tsareva P. E., , tsariovape@gmail.com, Saint-Petersburg State Marine Technical University, St. Petersburg, Russia
Minin M. S., , wasp-33@mail.ru, Saint-Petersburg State Marine Technical University, Saint-Petersburg, Russia
- Keywords
- computational fluid dynamics / parallel computing / GPU
- Year
- 2025 Issue 4 Pages 15 - 20
- Code EDN
- MFUHEU
- Code DOI
- 10.52190/1729-6552_2025_4_15
- Abstract
- Computational fluid dynamics problems are characterized by increased complexity and volume of calculations, so when developing software, the question of introducing a parallel structure is invariably raised in order to increase the efficiency of available resources and reduce the execution time of calculations. This article attempts to describe the main ways of implementing parallelism for general approaches to describing a fluid, approximating the solution, and the numerical methods used. The most universal methods are the transfer of calculations to a graphics card and geometric parallelism - the division of the calculation domain into subdomains, the calculation of values in which can be calculated simultaneously. Functional division of independent subtasks into different processors can also be implemented at the individual program level. The most promising approach is the division of the computational domain in time, which is the most difficult to implement. At the same time, in the context of maximum efficiency, any approaches should be considered individually for each task.
- Text
- BUY for read the full text of article
- Buy
- 500.00 rub
