Месяц Е.А.   Snytnikov A.V.  

The increase of PIC simulation accuracy for hybrid supercomputers using C++ templates

Reporter: Месяц Е.А.


E.A.Mesyats, A.V.Snytnikov

ICMMG SB RAS, Novosibirsk
snytav@ssd.sscc.ru, mesyats@gmail.com

The goal of the present work is the development of parallel numerical methods for the solution of some specific sort of plasma physics  problems. These problems are dominated by the plasma instabilities. These problems are to be studied within the kinetic approach [1]. It means the use of Particle-In-Cell (PIC) method. Thus the reduce of the numerical noise is the question of great importance. Moreover, the noise must be reduced without the particle number increase. Also, noise reduction and accuracy increase must be achieved without rewriting of the code, only small separate code parts being changed.

In order to reduce the noise level particle form-factors of different shape were being used. The shape was the B-spline of the 2 and 3 order. The form-factors are quickly and simply replaced in the code by the use of C++ templates with Esirkepov [2] exact charge conservation scheme for Particle-in-Cell simulation with an arbitrary form-factor. The simulation domain is implemented as a template class with the Cell class being template parameter. One of the Cell class methods  is  the  form-factor. In such a way, the use of higher order form-factors gives the more accurate simulation result, and the C++ templates facilitate the quick and easy form-factor replacement.

This work was supported by The Russian Science Foundation (Grant No. 14-12-00043).
Development of the code was supported by RFBR (Grants 15-31-20150, 14-01-00392 and 14-07-00241). Simulations are performed at Siberian Supercomputer Center SB RAS.

1. K.V.Lotov, I.V.Timofeev, E.A.Mesyats, A.V.Snytnikov, V.A.Vshivkov. Note on quantitatively correct simulations of the kinetic beam-plasma instability. // Physics of Plasmas, 22, 024502 (2015)
2. T.Zh. Esirkepov Exact charge conservation scheme for Particle-in-Cell simulation with an arbitrary form-factor // Computer Physics Communications.Computer Physics Communications. 2001. V. 135, iss. 2. P. 144-153.

To reports list