• DocumentCode
    3703925
  • Title

    Nonlinear analysis of the Lorenz system and buck converter using GPU

  • Author

    Guillermo Gallo;Reinel Tabarest;Gustavo Osorio

  • Author_Institution
    Departamento de Ingenier?a El?ctrica, Electr?nica y Computaci?n, Universidad Nacional de Colombia Sede Manizales, Kil?metro 7 v?a al Aeropuerto La Nubia. Campus La Nubia, Edificio Q, Piso 2, Oficina 205
  • fYear
    2015
  • Firstpage
    1
  • Lastpage
    6
  • Abstract
    In this paper, we present the bifurcation analysis and the atraction basins of two systems using brute force algorithms executed in parallel way on GPU. The algorithms were developed in C language and using CUDA libraries for the parallel programming. The first system to be analyzed is the Lorenz system. This smooth nonlinear system has been studied using GPU. The second one is a piecewise smooth system, this is the buck converter controlled by sine wave. There are not studies in the literature about nonlinear analysis of this system using GPU. Numerical experiments were performed in parallel way on two GPUs and sequentially on a CPU. We achieve speed ups near to 102X. Finally, we present some comparisons of execution times on different architectures.
  • Keywords
    "Graphics processing units","Kernel","Algorithm design and analysis","Random access memory","Bifurcation","Force","Libraries"
  • Publisher
    ieee
  • Conference_Titel
    Automatic Control (CCAC), 2015 IEEE 2nd Colombian Conference on
  • Type

    conf

  • DOI
    10.1109/CCAC.2015.7345227
  • Filename
    7345227