Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Parallelizing the fast multipole method for the Helmholtz

Conference ·
OSTI ID:125524
 [1]
  1. Hughes Research Labs., Malibu, CA (United States)
The fast multipole method (FMM) dramatically reduces the time required to compute radar cross sections and antenna radiation patterns compared to dense matrix techniques. The FMM decreases the operation count of the matrix-vector multiplication in iterative solvers to O(N{sup 3/2}), where N is the number of unknowns. Presented is a parallel FMM and the performance of its implementation on an Intel Paragon. For a 90,000 unknown problem running on 60 processors, the FMM representation fits in memory and the algorithm computes the matrix-vector product in 1.26 seconds. It sustains an aggregate rate of 1.4 Gflop/s. The corresponding dense matrix would occupy over 100 Gbytes and require approximately 50 seconds to form the product.
OSTI ID:
125524
Report Number(s):
CONF-950212--
Country of Publication:
United States
Language:
English

Similar Records

Applications of boundary element methods on the Intel Paragon
Conference · Mon Oct 31 23:00:00 EST 1994 · OSTI ID:10161914

Comparing precorrected-FFT and fast multipole algorithms for solving three-dimensional potential integral equations
Conference · Fri Dec 30 23:00:00 EST 1994 · OSTI ID:224493

An Adaptive Fast Multipole Boundary Element Method for Poisson-Boltzmann Electrostatics
Journal Article · Wed Dec 31 23:00:00 EST 2008 · Journal of Chemical Theory and Computation · OSTI ID:981440