Advertisement

The Hybrid-Cluster Multilevel Approach to Solving the Elastic Wave Propagation Problem

  • Boris Glinskiy
  • Anna Sapetina
  • Valeriy Martynov
  • Dmitry Weins
  • Igor ChernykhEmail author
Conference paper
Part of the Communications in Computer and Information Science book series (CCIS, volume 753)

Abstract

We propose in this paper an integrated approach to creating algorithms and software to simulate seismic problems. Our extended co-design concept considers supercomputer architecture at all stages, beginning with the physical and mathematical formulation of the problem, and then developing an algorithm and code. In particular, we compare the efficiency of the parallel implementation (on a supercomputer equipped with GPU) of algorithms for solving two mathematical statements of dynamic elasticity problem for the numerical modeling of seismic wave fields in 3D media, which is typical of volcanic structures.

The scalability of the algorithms is investigated using the multi-agent system AGNES to simulate the behavior of computational nodes based on the current state of computer equipment characteristics. We present the results obtained for the efficiency of the implementation of the algorithms when using millions of cores. Also, we assess the energy efficiency of these algorithms.

Keywords

Elastic waves 3D modeling Finite-difference schemes Hybrid cluster GPU Co-design Agent simulation Energy efficiency of algorithms 

References

  1. 1.
    Glinskii, B.M., Kovalevskii, V.V., Khairetdinov, M.S.: Vibroseismic monitoring of earthquake-prone areas. Volcanol. Seismol. 21(6), 723–730 (2000)Google Scholar
  2. 2.
    Alekseev, A.S., Glinsky, B.M., Kovalevsky, V.V., Khairetdinov, M.S.: Active vibromonitoring: experimental systems and fieldwork results. In: Handbook of Geophysical Exploration: Seismic Exploration, vol. 40, pp. 105–120 (2010). doi: 10.1016/s0950-1401(10)04011-5
  3. 3.
    Glinskii, B.M., Martynov, V.N., Sapetina, A.F.: 3D modeling of seismic wave fields in a medium specific to volcanic structures. Yakutian Math. J. 22(3), 84–98 (2015)zbMATHGoogle Scholar
  4. 4.
    Glinskiy, B.M., Kulikov, I.M., Snytnikov, A.V., Chernykh, I.G., Weins, D.: Mnogourovnevyj podhod k razrabotke algoritmicheskogo i programmnogo obespechenija jekzaflopsnyh superJeVM (A multilevel approach to algorithm and software design for exaflops supercomputers). Vychislitel’nye metody i programmirovanie (Vychisl. Metody Programm.) 16, 543–556 (2015)Google Scholar
  5. 5.
    Glinskiy, B., Kulikov, I., Snytnikov, A., Romanenko, A., Chernykh, I., Vshivkov, V.: Co-design of parallel numerical methods for plasma physics and astrophysics. Supercomput. Front. Innov. 1(3), 88–98 (2014). doi: 10.14529/jsfi140305 Google Scholar
  6. 6.
    Dosanjh, S.S., et al.: Exascale design space exploration and co-design. Future Gener. Comput. Syst. 30, 46–58 (2014). doi: 10.1016/j.future.2013.04.018 CrossRefGoogle Scholar
  7. 7.
    Bihn, M., Weiland, T.: A stable discretization scheme for the simulation of elastic waves. In: Proceedings of the 15th IMACS World Congress on Scientific Computation, Modelling and Applied Mathematics (IMACS 1997), Berlin, vol. 2, pp. 75–80 (1997)Google Scholar
  8. 8.
    Karavaev, D.A.: Parallel’naja realizacija metoda chislennogo modelirovanija volnovyh polej v trehmernyh modeljah neodnorodnyh sred (Parallel Implementation of Wave Field Numerical Modeling Method in 3D Models of Inhomogeneous Media). Vestnik Nizhegorodskogo universiteta im. N.I. Lobachevskogo (Vestnik of Lobachevsky State University of Nizhni Novgorod) 6(1), 203–209 (2009)Google Scholar
  9. 9.
    Nakata, N., Tsuji, T., Matsuoka, T.: Acceleration of computation speed for elastic wave simulation using a graphic processing unit. Explor. Geophys. 42(1), 98–104 (2011). doi: 10.1071/eg10039 CrossRefGoogle Scholar
  10. 10.
    Wooldridge, M.: Introduction to MultiAgent Systems. Wiley, England (2002)Google Scholar
  11. 11.
    Bellifemine, F.L., Caire, G., Greenwood, D.: Developing Multi-Agent Systems with JADE. Wiley, New York (2007). doi: 10.1002/9780470058411 CrossRefGoogle Scholar
  12. 12.
    Podkorytov, D., Rodionov, A., Sokolova, O., Yurgenson, A.: Using agent-oriented simulation system AGNES for evaluation of sensor networks. In: Vinel, A., Bellalta, B., Sacchi, C., Lyakhov, A., Telek, M., Oliver, M. (eds.) MACOM 2010. LNCS, vol. 6235, pp. 247–250. Springer, Heidelberg (2010). doi: 10.1007/978-3-642-15428-7_24 CrossRefGoogle Scholar
  13. 13.
    Podkorytov, D., Rodionov, A., Choo, H.: Agent-based simulation system AGNES for networks modeling: review and researching. In: Proceedings of the 6th International Conference on Ubiquitous Information Management and Communication (ACM ICUIMC 2012), p. 115. ACM (2012). ISBN 978-1-4503-1172-4. doi: 10.1145/2184751.2184883
  14. 14.
    Vins, D.V.: Analiz jeffektivnosti sistemy upravlenija potokom zadanij dlja CKP v mul’tiagentnoj imitacionnoj modeli (Analysis of Effectiveness of Job Stream Management System for the Center of Collective Use in Multi-agent Simulation Model). Vestnik NGU (Vestnik NSU: Information Technologies) 12(2), 33–41 (2014)Google Scholar
  15. 15.
    Glinsky, B.M., Marchenko, M.A., Mikhailenko, B.G., Rodionov, A.S., Chernykh, I.G., Karavaev, D.A., Podkorytov, D.I., Vins, D.V.: Simulation modeling of parallel algorithms for exaflop supercomputers (in Russian). Inf. Technol. Comput. Syst. 4, 3–14 (2013)Google Scholar
  16. 16.
    Kulikov, I., Chernykh, I., Glinsky, B., Weins, D., Shmelev, A.: Astrophysics simulation on RSC massively parallel architecture. In: Proceedings of 2015 IEEE/ACM 15th International Symposium on Cluster, Cloud, and Grid Computing, CCGrid, pp. 1131–1134. IEEE Press (2015). doi: 10.1109/ccgrid.2015.102

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Boris Glinskiy
    • 1
  • Anna Sapetina
    • 1
  • Valeriy Martynov
    • 1
  • Dmitry Weins
    • 1
  • Igor Chernykh
    • 1
    Email author
  1. 1.Institute of Computational Mathematics and Mathematical Geophysics of the Siberian Branch of the Russian Academy of SciencesNovosibirskRussia

Personalised recommendations