RUS  ENG
Full version
JOURNALS // Vestnik Udmurtskogo Universiteta. Matematika. Mekhanika. Komp'yuternye Nauki // Archive

Vestn. Udmurtsk. Univ. Mat. Mekh. Komp. Nauki, 2016 Volume 26, Issue 3, Pages 445–450 (Mi vuu551)

This article is cited in 5 papers

COMPUTER SCIENCE

The effectiveness of parallelizing an algorithm of the PFC equation solution using PetIGA library

I. O. Starodumova, E. V. Pavlyukb, S. M. Abramovc, L. V. Klyuevd, P. K. Galenkoe, D. V. Alexandrovf

a Laboratory of Multi-Scale Mathematical Modeling, Ural Federal University, pr. Lenina, 51, Yekaterinburg, 620000, Russia
b Department of Computational Mathematics, Ural Federal University, pr. Lenina, 51, Yekaterinburg, 620000, Russia
c Ailamazyan Program Systems Institute of the Russian Academy of Sciences, ul. Petra I, 4a, s. Ves'kovo, Pereslavskii raion, Yaroslavl oblast, 152021, Russia
d Immers Ltd., Michurinskii pr., 19, Moscow, 119192, Russia
e Faculty of Physics and Astronomy, Friedrich Schiller University, Jena, D-07743, Germany
f Department of Mathematical Physics, Laboratory of Multi-Scale Mathematical Modeling, Ural Federal University, pr. Lenina, 51, Yekaterinburg, 620000, Russia

Abstract: The paper presents an algorithm for solving the equation of Phase Field Crystal (PFC) in a hyperbolic statement that allows to describe the phase transitions of metastable or unstable state at the nuclear density scale, described by a differential equation of the sixth order with respect to the space variable and the second order with respect to the time variable. The algorithm is based on the method of isogeometric analysis (IGA) and is implemented by PetIGA library. The resulting code allows parallel computations, which significantly speeds up the process of solving a problem. The effectiveness of used instruments during the calculations on high-performance computing clusters is evaluated. An analysis of the effectiveness of the current algorithm is carried out for heterogeneous computer systems.

Keywords: phase field crystal, high performance computation, isogeometric analysis.

UDC: 519.711.3

MSC: 65D05

Received: 17.05.2016

DOI: 10.20537/vm160312



Bibliographic databases:


© Steklov Math. Inst. of RAS, 2024