Abstract
Due to a hard dependency between time steps, large-scale simulations of gas using the Direct Simulation Monte Carlo (DSMC) method proceed at the pace of the slowest processor. Scalability is therefore achievable only by ensuring that the work done each time step is as evenly apportioned among the processors as possible. Furthermore, as the simulated system evolves, the load shifts, and thus this load-balancing typically needs to be performed multiple times over the course of a simulation. Common methods generally use either crude performance models or processor-level timers. We combine both to create a timer-augmented cost function which both converges quickly and yields well-balanced processor decompositions. When compared to a particle-based performance model alone, our method achieves 2\(\times \) speedup at steady-state on up to 1024 processors for a test case consisting of a Mach 9 argon jet impacting a solid wall.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsNotes
- 1.
The names of the methods are our own labels for them.
References
Berger, M.J., Bokhari, S.H.: A partitioning strategy for nonuniform problems on multiprocessors. IEEE Trans. Comput. 5, 570–580 (1987)
Bird, G.A.: The DSMC Method. CreateSpace Independent Publishing Platform, Scotts Valley (2013)
Bird, G.A.: Molecular Gas Dynamics. NASA STI/Recon Technical Report A 76 (1976)
Boyd, I.D., Schwartzentruber, T.E.: Nonequilibrium Gas Dynamics and Molecular Simulation. Cambridge University Press, Cambridge (2017)
Ivanov, M., Markelov, G., Taylor, S., Watts, J.: Parallel DSMC strategies for 3D computations. North-Holland, Amsterdam (1997). https://doi.org/10.1016/B978-044482327-4/50128-5, http://www.sciencedirect.com/science/article/pii/B9780444823274501285
Jülich Supercomputing Centre: JURECA: general-purpose supercomputer at Jülich supercomputing centre. J. Large-Scale Res. Facil. 2(A62) (2016). https://doi.org/10.17815/jlsrf-2-121, http://dx.doi.org/10.17815/jlsrf-2-121
Karniadakis, G.E., Beskok, A., Aluru, N.: Microflows and Nanoflows: Fundamentals and Simulation, vol. 29. Springer, New York (2006). https://doi.org/10.1007/0-387-28676-4
McDoniel, W.J., Goldstein, D.B., Varghese, P.L., Trafton, L.M.: Three-dimensional simulation of gas and dust in Io’s Pele plume. Icarus 257, 251–274 (2015). https://doi.org/10.1016/j.icarus.2015.03.019. http://www.sciencedirect.com/science/article/pii/S0019103515001190
McDoniel, W.J., Goldstein, D.B., Varghese, P.L., Trafton, L.M.: The interaction of Io’s plumes and sublimation atmosphere. Icarus 294, 81–97 (2017). https://doi.org/10.1016/j.icarus.2017.04.021. http://www.sciencedirect.com/science/article/pii/S0019103517303068
Moss, J.N., Bird, G.A.: Direct simulation of transitional flow for hypersonic re-entry conditions. Therm. Desi. Aeroassisted Orbital Transf. Veh. 96, 338–360 (1985)
Nance, R.P., Wilmoth, R.G., Moon, B., Hassan, H., Saltz, J.: Parallel DSMC solution of three-dimensional flow over a finite flat plate. NASA Technical report (1994)
Olson, S.E., Christlieb, A.J., Fatemi, F.K.: PID feedback for load-balanced parallel gridless DSMC. Comput. Phys. Commun. 181(12), 2063–2071 (2010). https://doi.org/10.1016/j.cpc.2010.06.045. http://www.sciencedirect.com/science/article/pii/S0010465510002985
Taylor, S., Watts, J.R., Rieffel, M.A., Palmer, M.E.: The concurrent graph: basic technology for irregular problems. IEEE Parallel Distrib. Technol.: Syst. Appl. 4(2), 15–25 (1996)
Weinberg, M.D.: Direct simulation monte carlo for astrophysical flows-II. Ram-pressure dynamics. Mon. Not. R. Astron. Soc. 438(4), 3007–3023 (2014)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this paper
Cite this paper
McDoniel, W., Bientinesi, P. (2019). A Timer-Augmented Cost Function for Load Balanced DSMC. In: Senger, H., et al. High Performance Computing for Computational Science – VECPAR 2018. VECPAR 2018. Lecture Notes in Computer Science(), vol 11333. Springer, Cham. https://doi.org/10.1007/978-3-030-15996-2_12
Download citation
DOI: https://doi.org/10.1007/978-3-030-15996-2_12
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-15995-5
Online ISBN: 978-3-030-15996-2
eBook Packages: Computer ScienceComputer Science (R0)