Accelerating Multi GPU Based Discontinuous Galerkin FEM Computations for Electromagnetic Radio Frequency Problems
Keywords:
Accelerating Multi GPU Based Discontinuous Galerkin FEM Computations for Electromagnetic Radio Frequency ProblemsAbstract
A Graphics Processing Unit (GPU) accelerated simulation of Maxwell’s equations in the time domain is presented. The Discontinuous Galerkin Finite Element Method (DG-FEM) is used for discretization since the elementwise structure fits the parallelization design aspects of the GPU architecture and the NVIDIA Compute Unified Device Architecture (CUDA), a GPU programming model. The parallelization strategy for a multi-GPU setup using CUDA is focused. Several performance improvements are analyzed and investigated with the help of a realistic 3D electromagnetic scattering example.
Downloads
References
Nvidia Corporation. “NVIDIA
Compute Unified Device
Programming Guide”, USA, 2009.
Advanced Micro Devices, Inc., “
Computing”, Sunnyvale, USA, 200
KHRONOS GROUP, “The
Specification Version 1.0”, 2008.
S. Krakiwsky, L. Turner, and M.
“Acceleration of Finite-Differe
Domain (FDTD) Using Graphic
Units(GPU)”, IEEE MTT-S I
Microwave Symposium, pp. 1033-1
M. J. Inman, A. Z. Elsherbeni, J.
and B. N. Baker, “Practical Implem
a CPML Absorbing Boundary
Accelerated FDTD Techniqu
Journal, vol. 23, 2008.
A. Kloeckner, T.Warburton, J. B r
Hesthaven, “Nodal discontinuou
methods on graphics processors,"
Computational Physics, vol. 228 ,
, 2009.
N. Gödel, T. Warburton, M. Cle m
Accelerated Discontinuous Galerk
Electromagnetic Radio Frequency
IEEE APS Conference Charleston,
W. Reed and T. Hill, “Trian
methods for the neutron transpor
Los Alamos Scientific Laboratory
Report, no. LAUR-73-479, 1973.
B. Cockburn, G. Karniadakis, and
“Discontinuous Galerkin Metho
Computation and Applications”
G. Cohen, X. Ferrieres and S.
spatial high-order hexahedral d
Galerkin method to solve Maxwel
in time domain”, Journal of C o
Physics, vol. 217, pp. 340-363, 200
J. S. Hesthaven and T.
“Discontinuous Galerkin metho
time-domain Maxwell’s equatio
Journal, vol. 19, pp. 10–29, 2004.
J. S. Hesthaven and T. Warbu
Discontinuous Galerkin Methods.
G. Karypis and V. Kumar, “A f
quality multilevel scheme for
irregular graphs." Conference
Processing, pp. 113-122, 1995.
CUDA 2.2
Architecture
.
“ATI Stream
OpenCL
Okoniewski,
ence Time-
cs Processor
International
, 2004.
G. Maloney
mentation of
y for GPU
ue”, ACES
ridge, and J.
us Galerkin
" Journal of
, pp. 7863 –
mens, “GPU
kin FEM for
y Problems”,
, 2009.
ngular mesh
rt equation,”
y, vol. Tech.
d C.-W. Shu,
ods: Theory,
”, Springer,
Pernet, “A
discontinuous
ll's equations
omputational
Warburton,
ods fo r the
ons,” ACES
urton, Nodal
Springer,
fast and high
partitioning
on Parallel
