Using Modern Multi-/Many-core Architecture for the Engineering Simulations

Chapter

Abstract

The paper presents authorial method for computer simulation of the process of the casting solidification process using modern multi-/many-core architectures. The authors have focused above all on the GPU compatible with CUDA architecture. Presented in the article method enables division the process of the building a global matrix of coefficients in two independent phases. The first one is performed only once before the start of a computer engineering simulation, the second phase is performed at each time step of computer simulation. The reason for such division is decoupling of computations realised on the GPU from the spatial informations (e.g. physical coordinates of the nodes) contained in the mesh of finite elements. Application of presented method enables the computer scientific simulation in a variety of multicore architectures, such as multi-core CPUs, GPUs, APUs and Intel Phi. The use of computer engineering simulations multi-core processor allows significant acceleration computation, thereby also reduces costs. Presented in the work results have been obtained by using authorial software built by authors from scratch.

Keywords

Casting Computer simulation CUDA Enthalpy Graphics processor High performance computing Parallel processing Solidification processing 

References

  1. 1.
    H.T. Meng, B.L. Nie, S. Wong, C. Macon, J.M. Jin, GPU accelerated finite-element computation for electromagnetic analysis. IEEE Antennas Propag. Mag. 56, 39–62 (2014)CrossRefGoogle Scholar
  2. 2.
    G. Michalski, N. Sczygiol, Using CUDA architecture for the computer simulation of the casting solidification process. Lect Notes Eng. Comp.: Proc. Int. MultiConf. Eng. Comput. Sci. 2, 933–937 (2014)Google Scholar
  3. 3.
    R. Strzodka, M. Dogger, A. Kolb, Scientific computation for simulations on programmable graphics hardware. Simul. Model. Pract. Theory 13, 667–680 (2005)CrossRefGoogle Scholar
  4. 4.
    S. Che, M. Boyer, J. Meng, D. Tarjan, J.W. Sheaffer, K. Skadron, A performance study of general-purpose applications on graphics processors using CUDA. J. Parallel Distrb. Com. 68, 1370–1380 (2008)CrossRefGoogle Scholar
  5. 5.
    nVidia: CUDA C best practices guide. Design guide v. 6.0 (2014). http://docs.nvidia.com/cuda/#axzz398vseVcQ Cited 01 Aug 2014
  6. 6.
    P. Pospíchal, J. Schwarz, J. Jaroš, Parallel genetic algorithm solving 0/1 knapsack problem running on the GPU, in 16th International Conference on Soft Computing MENDEL 2010. pp. 64–70Google Scholar
  7. 7.
    Ch. Lee, X. Wei, J.W. Kysar, J. Hone, Measurement of the elastic properties and intrinsic strength of monolayer graphene. Science 321, 385–388 (2008)CrossRefGoogle Scholar
  8. 8.
    N. Sczygiol, Numerical modelling of thermo-mechanical phenomena in a solidifying casting and mold (Wyd PCz, Czestochowa, 2000). (in Polish)Google Scholar
  9. 9.
    N. Sczygiol, G. Szwarc, Application of enthalpy formulations for numerical simulation of castings solidification. Comput. Assist. Mech. Eng. Sci. 8, 99–120 (2001)MATHGoogle Scholar
  10. 10.
    O. Zienkiewicz, The finite element method, vol I, the Basis, 5th ed. Oxford: Butterworth-HeinemannGoogle Scholar
  11. 11.
    N. Sczygiol, G. Szwarc, R. Wyrzykowski, Numerical modelling of equiaxed structure formation during solidification of a two-component alloy, in Second European Conference on Computational Mechanics, Kraków, pp. 820–821Google Scholar
  12. 12.
    E. Gawronska, N. Sczygiol, Application of mixed time partitioning methods to raise the efficiency of solidification modeling, in 12th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC), doi:http://dx.doi.org/10.1109/SYNASC.2010.24
  13. 13.
    E. Gawronska, N. Sczygiol, Relationship between eigenvalues and size of time step in computer simulation of thermomechanics phenomena. Lect. Notes Eng. Comp.: Proc. Int. MultiConf. Eng. Comput. Sci. 2, 881–885 (2014)Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2015

Authors and Affiliations

  1. 1.Institute of Computer and Information SciencesCzestochowa University of TechnologyCzestochowaPoland

Personalised recommendations