Study and analysis of parallel N-body simulations

• Guilherme G. Arcencio UFSCar
• Hélio Crestana Guardia UFSCar

DOI: https://doi.org/10.5753/eradsp.2024.239679

Resumo

The N-body Problem involves predicting the motion of a group of astronomic particles under the influence of each other’s gravitational field. Due to the nonexistence of closed-form analytical solutions for N bigger than 2, numerical methods must be employed, and those do not scale well for large numbers of bodies. Thus, we developed and investigated parallel, high performance N-body simulations in GPUs and multi-core CPUs, and found that, even with dynamic time steps and synchronization requirements, GPUs massively reduce the time required by the simulations. While multi-core CPUs also allow for large speedups, they are vastly outperformed by the former.

Referências

Barnes, J. and Hut, P. (1986). A hierarchical O(N log N) force-calculation algorithm. Nature, 324(6096):446–449.

Board, J. A., Humphres, C. W., Lambert, C. G., Rankin, W. T., and Toukmaji, A. Y. (1999). Ewald and multipole methods for periodic N-body problems. In Computational Molecular Dynamics: Challenges, Methods, Ideas, pages 459–471, Berlin, Heidelberg. Springer.

Boukhary, S. and Colmenares, E. (2019). Study, analysis, and acceleration of an n-body simulation under many-core environments using an object oriented approach. In 2019 International Conference on Computational Science and Computational Intelligence, pages 1506–1510. IEEE.

Greengard, L. and Rokhlin, V. (1987). A fast algorithm for particle simulations. Journal of Computational Physics, 73(2):325–348.

Grudić, M. Y. and Hopkins, P. F. (2020). A general-purpose time-step criterion for simulations with gravity. Monthly Notices of the Royal Astronomical Society, 495:4306–4313.

Hairer, E., Lubich, C., and Wanner, G. (2003). Geometric numerical integration illustrated by the Störmer–Verlet method. Acta Numerica, 12:399–450.

Meyer, K., Hall, G., and Offin, D. (2009). Introduction to Hamiltonian Dynamical Systems and the N-Body Problem. Springer.

Pinto, V. G., Herbstrith, V. A., and Schnorr, L. M. (2015). Replicating the performance evaluation of an n-body application on a manycore accelerator. In 2015 International Symposium on Computer Architecture and High Performance Computing Workshop, pages 19–24. IEEE.

Zecena, I., Burtscher, M., Jin, T., and Zong, Z. (2013). Evaluating the performance and energy efficiency of n-body codes on multi-core CPUs and GPUs. In 2013 IEEE 32nd International Performance Computing and Communications Conference, pages 1–8. IEEE.