Study on the Particle Sorting Performance for Reactor Monte Carlo Neutron Transport on Apple Unified Memory GPUs (2401.11455v2)
Abstract: In simulation of nuclear reactor physics using the Monte Carlo neutron transport method on GPUs, the sorting of particles plays a significant role in performance of calculation. Traditionally, CPUs and GPUs are separated devices connected at low data transfer rate and high data transfer latency. Emerging computing chips tend to integrate CPUs and GPUs. One example is the Apple silicon chips with unified memory. Such unified memory chips have opened doors for new strategies of collaboration between CPUs and GPUs for Monte Carlo neutron transport. Sorting particle on CPU and transport on GPU is an example of such new strategy, which has been suffering the high CPU-GPU data transfer latency on the traditional devices with separated CPU and GPU. The finding is that for the Apple M2 max chip, sorting on CPU leads to better performance per power than sorting on GPU for the ExaSMR whole core benchmark problems and the HTR-10 high temperature gas reactor fuel pebble problem. The partially sorted particle order has been identified to contribute to the higher performance with CPU sort than GPU. The in-house code using both CPU and GPU achieves 7.5 times power efficiency that of OpenMC on CPU for ExaSMR whole core benchmark with depleted fuel, and 150 times for HTR-10 fuel pebble benchmark with depleted fuel.
- S. Hamilton and T. Evans, “Continuous-energy Monte Carlo neutron transport on GPUs in the Shift code,” Annuals of Nuclear Energy, vol. 128, pp. 236-247, 2019.
- N. Choi and H. Joo, “Domain decomposition for GPU-based continuous energy Monte Carlo power reactor calculation,” Nuclear Engineering and Technology, vol. 52, issue 11, pp. 2667-2677, 2020.
- K. Gao, Z. Chen, A. Sun and T. Yu, “The research and application of GPU-based Monte Carlo Simulation in reactor calculation,” Proceedings of RPNM2023, Jul. 26-29 Lanzhou China, 2023.
- J. Tramm, P. Romano, J. Doerfert, A. Lund, P. Shriwise, A. Siegel, and et. al., “Toward Portable GPU Acceleration of the OpenMC Monte Carlo Particle Transport Code,” Proceedings of PHYSOR2022, May 15-20 Pittsburg USA. 2022.
- R. Bergmman, J. Vujić, “Algorithmic choices in WARP - A framework for continuous energy Monte Carlo neutron transport in general 3D geometries on GPUs,” Annuals of Nuclear Energy, vol. 77, pp. 176–193, 2015.
- C. Liu, “Doppler broadening using discrete cosine transform and kernel reconstruction for spatially variable media,” Annuals of Nuclear Energy, vol. 174, pp. 109150, 2012.
- P. Romano, J. Tramm, P. Shriwise, “Impact of Asymmetric Multicore Processors on Monte Carlo Particle Transport Code Performance,” Proceedings of M&C 2023 (394), Aug. 13-17 Niagara Falls Canada, 2023.
- J. Tramm, K. Yoshii, P. Romano, “Power at Your Fingertips: Assessing the Performance of a Monte Carlo Neutron Transport Mini-App on Consumer Laptop GPUs,” Proceedings of M&C 2023 (433), Aug. 13-17 Niagara Falls Canada, 2023.
- C. Liu, “Monte Carlo neutron transport using low power mobile GPU devices”, Arxiv, https://arxiv.org/abs/2208.06296, 2022
- B. Godfrey, “VERA core physics benchmark progression problem specifications, revision 4,” CASL technical report CASL-U-2012-0131-004, 2014.
- E. Merzari, S. Hamilton, T. Evans, M. Min and et. al., “Exascale Multiphysics Nuclear Reactor Simulations for Advanced Designs,” Proceedings of SC23, Nov. 12-17 Denver USA, https://doi.org/10.1145/3581784.3627038, 2023
- International Handbook of Reactor Physics Experiments, “Evaluation of the Initial Critical Configuration of the HTR-10 Pebble-Bed Reactor,” HTR10-GCR-RESR-001, NEA/NSC/DOC(2006)1, Rev. 0., 2006
- Y. Cheng, C. Hao and F. Li, “Uncertainty quantification of fuel pebble model and its effect on the uncertainty propagation of nuclear data in pebble bed HTR,” Annuals of Nuclear Energy, vol. 139, pp. 107286, 2020.