Optimization of Numerical Simulation of a Plasma with a Macroparticle using OpenMP and MPI

In this work, the program developed in C language was optimized for numerically simulating the wake potential at various parameters of the plasma flow and the algorithm was adapted for execution on cluster systems with shared and distributed memory. The computational complexity of the algorithm and the required amount of RAM does not allow finding an analytical solution for the practically important sizes of the numerical model. This work aims to accelerate the implementation of a numerical solution of a physical problem using the example of modeling a plasma with a particulate and to develop a methodology for optimizing the numerical solution of a physical problem using OpenMP and MPI technologies. Computing experiments carried out on the HSE supercomputer have shown that using OpenMP technology can reduce program execution time, and calculations on distributed memory systems can significantly reduce the amount of RAM on each node. Particular attention is paid to optimizing single-threaded code, as well as working with the above technologies. Also in the course of the work, methods for optimizing this and similar tasks were formulated.