Article

The Journal of Supercomputing

, Volume 63, Issue 3, pp 897-918

Open Access This content is freely available online to anyone, anywhere at any time.

Adaptive fast multipole methods on the GPU

  • Anders GoudeAffiliated withDivision of Electricity, Department of Engineering Sciences, Uppsala University
  • , Stefan EngblomAffiliated withDivision of Scientific Computing, Department of Information Technology, Uppsala University Email author 

Abstract

We present a highly general implementation of fast multipole methods on graphics processing units (GPUs). Our two-dimensional double precision code features an asymmetric type of adaptive space discretization leading to a particularly elegant and flexible implementation. All steps of the multipole algorithm are efficiently performed on the GPU, including the initial phase, which assembles the topological information of the input data. Through careful timing experiments, we investigate the effects of the various peculiarities of the GPU architecture.

Keywords

Adaptive fast multipole method CUDA Graphics processing units Tesla C2075