The Integrated Approach to Solving Large-Size Physical Problems on Supercomputers

This paper presents the results obtained by the authors on applying an integrated approach to solving geoseismics, astrophysics, and plasma physics problems on high-performance computers. The concept of the integrated approach in the context of mathematical modeling of physical processes is understood as constructing a physico-mathematical model of a phenomenon, a numerical method, a parallel algorithm and its software implementation with the efficient use of a supercomputer architecture. With this approach, it becomes relevant to compare not only the methods of solving a problem but, also, physical and mathematical statements of a problem aimed at creating the most effective implementation of a chosen computing architecture. The scalability of algorithms is investigated using the multi-agent system AGNES simulating the behavior of computing nodes based on the current state of computer equipment characteristics. In addition, special attention in this paper is given to the energy efficiency of algorithms.

[1]  Igor Kulikov,et al.  GPUPEGAS: A NEW GPU-ACCELERATED HYDRODYNAMIC CODE FOR NUMERICAL SIMULATIONS OF INTERACTING GALAXIES , 2014 .

[2]  Agostino Poggi,et al.  Developing Multi-agent Systems with JADE , 2007, ATAL.

[3]  Srivaths Ravi,et al.  Efficient RTL power estimation for large designs , 2003, 16th International Conference on VLSI Design, 2003. Proceedings..

[4]  Michael Wooldridge,et al.  Introduction to multiagent systems , 2001 .

[5]  Thomas L. Sterling,et al.  Achieving scalability in the presence of Asynchrony for Exascale Computing , 2012, High Performance Computing Workshop.

[6]  Jack J. Dongarra,et al.  Exascale computing and big data , 2015, Commun. ACM.

[7]  M. V. Popov,et al.  Piecewise parabolic method on a local stencil for ideal magnetohydrodynamics , 2008 .

[8]  Kirk W. Cameron,et al.  E-AMOM: an energy-aware modeling and optimization methodology for scientific applications , 2014, Computer Science - Research and Development.

[9]  James Demmel,et al.  the Parallel Computing Landscape , 2022 .

[10]  Da Qi Ren,et al.  Algorithm level power efficiency optimization for CPU-GPU processing element in data intensive SIMD/SPMD computing , 2011, J. Parallel Distributed Comput..

[11]  Hyunseung Choo,et al.  Agent-based simulation system AGNES* for networks modeling: review and researching , 2012, ICUIMC.

[12]  Rossiĭskai︠a︡ akademii︠a︡ nauk.  Computational mathematics and mathematical physics , 1992 .

[13]  Fabio Bellifemine,et al.  Developing Multi-Agent Systems with JADE (Wiley Series in Agent Technology) , 2007 .

[14]  Austria,et al.  Stellar hydrodynamical modeling of dwarf galaxies: simulation methodology, tests, and first results , 2015, 1504.04454.

[15]  David E. Keyes,et al.  Exaflop/s: The why and the how , 2011 .

[16]  Boris Glinskiy,et al.  A multilevel approach to algorithm and software design for exaflops supercomputers , 2015 .

[17]  Igor M. Kulikov,et al.  Using the PPML approach for constructing a low-dissipation, operator-splitting scheme for numerical simulations of hydrodynamic flows , 2016, J. Comput. Phys..

[18]  Russia,et al.  Collisionless stellar hydrodynamics as an efficient alternative to N-body methods , 2012, 1210.5246.

[19]  V. A. Vshivkov,et al.  HYDRODYNAMICAL CODE FOR NUMERICAL SIMULATION OF THE GAS COMPONENTS OF COLLIDING GALAXIES , 2011 .

[20]  P. Huynh,et al.  HERACLES: a three-dimensional radiation hydrodynamics code , 2007 .

[21]  Bronis R. de Supinski,et al.  Adagio: making DVS practical for complex HPC applications , 2009, ICS.