POSIX threads parallelization for example of Particle-In-Cell density calculations in plasma computer simulations

Anna Sasak, Marcin Brzuszek

Abstract


The TRQR program [1-4] simulates trajectories of charged particles (electrons or ions) in the electromagnetic field. TRQR is based on the Particle-In-Cell method whose basic guideline is the use of computational particles (called macro particles) that represent a large number of real particles of the same kind moving in the same direction. The program calculates particles charge density distribution and potential distribution for chosen ion sources, analyses particles behaviour in the electromagnetic field, describes the process of beams from the source extraction. A number of factors influences simulation results. In order to improve efficiency the program has been parallelized. This paper presents the process of converting chosen parts of the TRQR program into the multi-thread version. In the first step the program was moved from Fortran 77 to C++. Then it was parallelized using the Pthread library with the standard API for C++ contained in the POSIX IEEE 1003.1c standard. Each of threads has its own stack, set of registers, program counter, individual data, local variables, state information. All threads of particular process share one address space, general signal operations, virtual memory, data, input and output. The Mutex functions were used as a synchronization mechanism. This paper presents the analysis of a particular piece of main program that implements computations of particles density distribution. The paper presents execution time dependencies for different simulation parameters such as: the number of macro particles, size of the simulation mesh and the number of used threads.

Full Text:

PDF


DOI: http://dx.doi.org/10.2478/v10065-010-0050-8
Date of publication: 2010-01-01 00:00:00
Date of submission: 2016-04-27 16:26:34


Statistics


Total abstract view - 397
Downloads (from 2020-06-17) - PDF - 0

Indicators



Refbacks

  • There are currently no refbacks.


Copyright (c) 2015 Annales UMCS Sectio AI Informatica

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.