Главная страница
russian   english
16+
<< back

Title of Article

FORMING MATRICES OF FINITE ELEMENT SYSTEMS IN GPGPU


Issue
3
Date
2014

Article type
scientific article
UDC
004.272.2: 519.63
Pages
120-125
Keywords
parallel algorithms, numerical integration of finite element matrices, CUDA, computing on GPUs


Authors
Novikov A.K.
Institut mekhaniki UrO RAN, Izhevsk

Kopysov S.P.
Institut mekhaniki UrO RAN, Izhevsk

Kuzmin I.M.
Institut mekhaniki UrO RAN, Izhevsk

Nedozhogin N.S.
Institut mekhaniki UrO RAN, Izhevsk


Abstract
Parallel algorithms for forming matrices of finite element systems running on GPUs are considered. The features of the proposed algorithms in CUDA technology are discussed. The results of numerical simulation of continuum mechanics problems on hybrid computing nodes are presented.

File (in Russian)