Dynamic Load Balancing Based on Rectilinear Partitioning in Particle-in-Cell Plasma Simulation
Paper i proceeding, 2015

This paper considers load balancing in Particle-in-Cell plasma simulation on cluster systems. We propose a dynamic load balancing scheme based on rectilinear partitioning and discuss implementation of efficient imbalance estimation and rebalancing. We analyze the impact of load balancing on performance and accuracy. On a test plasma heating problem dynamic load balancing yields nearly 2 times speedup and better scaling. On the real-world plasma target irradiation simulation load balancing allows to mitigate particle resampling and thus improve accuracy of the simulation without increasing the runtime. Balancing-related overhead in both cases are under 1.5% of total run time.

Load balancing

Particle-in-cell

Plasma simulation

High performance computing

Författare

I. Surmin

Lobachevsky State University of Nizhni Novgorod

A. Bashinov

Lobachevsky State University of Nizhni Novgorod

Russian Academy of Sciences

S. Bastrakov

Lobachevsky State University of Nizhni Novgorod

E. Efimenko

Lobachevsky State University of Nizhni Novgorod

Russian Academy of Sciences

Arkady Gonoskov

Chalmers, Teknisk fysik, Kondenserade materiens teori

I. Meyerov

Lobachevsky State University of Nizhni Novgorod

Lecture Notes in Computer Science

0302-9743 (ISSN)

Vol. 9251 107-119

Ämneskategorier

Programvaruteknik

DOI

10.1007/978-3-319-21909-7_12

ISBN

978-3-319-21909-7