Using Parallel Computing in Computationally Intensive Problems of Simulating Particle Motion and Interaction in Plasma during Carbon Nanostructure Synthesis

Modern parallel computing technology makes it possible to solve complex computationally intensive problems of simulating physical processes using a personal computer. The paper considers a numerical solution to the equations forming a mathematical model of multicomponent plasma particle motion and interaction during arc discharge synthesis of carbon nanostructures. A large number of interacting particles (10¹⁶...10¹⁷) to be taken into account at each time step means considerable computational expense and time. Conventional numerical simulations of this type involve supercomputers or cloud computing. However, parallel computing technologies such as GPGPU and Nvidia CUDA make it possible to perform general purpose computations on graphics processing unit cores. We present software algorithms based on a modified particle-in-cell method which enable numerical simulations of the model under consideration to be run on a personal computer in a reasonable timeframe. We propose approaches to decomposing computation problems accounting for result synchronisation. We list those difficulties that arise while implementing collision detection algorithms for particles in plasma in actual software code and propose ways of overcoming these difficulties. We calculated the efficiency of various parallel algorithms

References

[1] Abramov G.V., Mironchenko E.A., Tolstova I.S. Mathematical modeling of processes in the electric arc synthesis of carbon nanotubes. Vestnik VGUIT [Proceedings of the Voronezh State University of Engineering Technologies], 2013, no. 4, pp. 106–110 (in Russ.).

[2] Tsvetkov I.V. Primenenie chislennykh metodov dlya modelirovaniya protsessov v plazme [Application of numerical methods for simulation plasma processes]. Moscow, MEPhI Publ., 2007. 84 p.

[3] Ispolzovanie videokart dlya vychisleniy [Using video-card for calculations]. GPGPU.ru: website (in Russ.). Available at: http://www.gpgpu.ru (accessed: 25.11.2017).

[4] Nvidia CUDA — negraficheskie vychisleniya na graficheskikh protsessorakh [Nvidia CUDA: Non-graphic calculations on graphic processors]. ixbt.com: website (in Russ.). Available at: http://www.ixbt.com/video3/cuda-1.shtml (accessed: 25.11.2017).

[5] Abramov G.V., Gavrilov A.N. Mathematical modeling of motion of interacting particles on the basis of the distribution functions in the plasma arc synthesis of ONS. Vestnik VGUIT [Proceedings of the Voronezh State University of Engineering Technologies], 2012, no. 2, pp. 71–75 (in Russ.).

[6] Cercignani C. Theory and application of the Boltzmann equation. Chatto and Windus, 1975. 415 p.

[7] Abramov G., Gavrilov A., Tolstova I. The use of technology for parallelization method of large particles using cloud computing. British Journal of Science, Education and Culture, 2014, no. 2 (6), pp. 380–387.

[8] Graphics add-in board shipments decline 15.2 % from last quarter. Available at: https://www.jonpeddie.com/press-releases/graphics-add-in-board-shipments-decline-15.2-from-last-quarter (accessed: 25.11.2017).

[9] Demmel J.W. Applied numerical linear algebra. SIAM, 1997. 184 p.

[10] Uskov R.V. On some peculiarities of application of CUDA technology for simulation of transport of radiation. Vestn. Mosk. Gos. Tekh. Univ. im. N.E. Baumana, Estestv. Nauki [Herald of the Bauman Moscow State Tech. Univ., Nat. Sci.], 2011, no. 3, pp. 71–83 (in Russ.).

[11] Baraff D. Dynamic simulation of non-penetrating rigid bodies. Ph. D. thesis. Cornell University, Computer Science Department, 1992, pp. 52–56.