Abstract
This paper considers Monte Carlo photon transport applications on heterogenous compute architectures with both CPUs and GPUs. Previous work on this problem has considered only meshes that can fully fit within the memory of a GPU, which is a significant limitation: many important problems require meshes that exceed memory size. We address this gap by introducing a new dynamic replication algorithm that adapts assignments based on the computational ability of a resource. We then demonstrate our algorithm’s efficacy on a variety of workloads, and find that incorporating the CPUs provides speedups of up to 20% over the GPUs alone. Further, these speedups are well beyond the FLOPS contribution from the CPUs, which provide further justification for continuing to include CPUs even when powerful GPUs are available. In all, the contribution of this work is an algorithm that can be applied in real-world settings to make more efficient use of heterogeneous architectures.
NOTICE: This manuscript has been authored by Lawrence Livermore National Security, LLC under Contract No. DE-AC52-07NA2 734-I with the US. Department of Energy. The United States Government retains, and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. LLNL-CONF-817536
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Livermore computing center high performance computing: Rzansel. https://hpc.llnl.gov/hardware/platforms/rzansel. Accessed 9 Dec 2020
Alme, H.J., Rodrigue, G.H., Zimmerman, G.B.: Domain decomposition models for parallel Monte Carlo transport. J. Supercomputing 18(1), 5–23 (2001)
Brantley, P., et al.: A new implicit Monte Carlo thermal photon transport capability developed using shared Monte Carlo infrastructure. In: The International Conference on Mathematics and Computational Methods Applied to Nuclear Science and Engineering (M and C 2019), Portland, Oregon, pp. 25–29. (2019)
Brunner, T.A., Brantley, P.S.: An efficient, robust, domain-decomposition algorithm for particle Monte Carlo. J. Comput. Phys. 228(10), 3882–3890 (2009)
Brunner, T.A., et al.: Comparison of four parallel algorithms for domain decomposed implicit Monte Carlo. J. Comput. Phys. 212(2), 527–539 (2006)
Ellis, J.A., et al.: Optimization of processor allocation for domain decomposed Monte Carlo calculations. Parallel Comput. 87, 77–86 (2019)
Horelik, N., Siegel, A., Forget, B., Smith, K.: Monte Carlo domain decomposition for robust nuclear reactor analysis. Parallel Comput. 40(10), 646–660 (2014)
O’Brien, M., et al.: Hybrid CPU-GPU load balancing for Monte Carlo particle transport. In: Proceedings of the 26th International Conference on Transport Theory (ICTT-26), Sorbonne Univeristy, Paris, France (2019)
O’Brien, M.J., Brantley, P.S., Joy, K.I.: Scalable load balancing for massively parallel distributed Monte Carlo particle transport. In: Proceedings of International Conference on Mathematics and Computational Methods Applied to Nuclear Science and Engineering (M and C 2013), Sun Valley, Idaho, vol. 45, pp. 647–658 (2013)
O’Brien, M.: Dynamic load balancing of parallel Monte Carlo transport calculations via spatial redecomposition. In: Proceedings of the Joint International Topical Meeting on Mathematics and Computation and Supercomputing in Nuclear Applications, pp. 16–19 (2007)
Procassini, R., O’Brien, M., Taylor, J.: Load balancing of parallel Monte Carlo transport calculations. In: Proceedings of the 2005 ANS Topical Meeting in Mathematics and Computation, Avignon, France, 12–15 September 2005 (2005)
Romano, P.K.: Parallel algorithms for Monte Carlo particle transport simulation on exascale computing architectures. Ph.D. thesis, Massachusetts Institute of Technology (2013)
Wagner, J.C., et al.: Hybrid and parallel domain-decomposition methods development to enable Monte Carlo for reactor analyses. Prog. Nucl. Sci. Technol. 2(1), 815–820 (2011)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Bleile, R., Brantley, P., O’Brien, M., Childs, H. (2021). A Dynamic Replication Approach for Monte Carlo Photon Transport on Heterogeneous Architectures. In: Paszynski, M., Kranzlmüller, D., Krzhizhanovskaya, V.V., Dongarra, J.J., Sloot, P.M.A. (eds) Computational Science – ICCS 2021. ICCS 2021. Lecture Notes in Computer Science(), vol 12742. Springer, Cham. https://doi.org/10.1007/978-3-030-77961-0_20
Download citation
DOI: https://doi.org/10.1007/978-3-030-77961-0_20
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-77960-3
Online ISBN: 978-3-030-77961-0
eBook Packages: Computer ScienceComputer Science (R0)