Computational Science Technical Note CSTN-120

CSTN Home

Numerical Precision and Benchmarking Very-High-Order Integration of Particle Dynamics on GPU Accelerators

K.A. Hawick, D.P. Playne and M.G.B. Johnson

Archived February 2011, Revised, April 2011

Abstract

GPUs offer a powerful acceleration platform for many scientific applications. Numerical integration of classical Newtonian dynamical particles often requires very high-order numerical accuracy. We assess the floating-point precision and performance of various GPUs for applications involving high-order time-step integration methods for particle model simulations using N-squared interactions. We demonstrate how high-order algorithms can be expressed in Compute Unified Device Architecture (CUDA) and present some detailed benchmark data. We show the high numerical power of high-order integration methods such as Hairer's $10^{th}$ order method and relate its performance to its precision requirements.

Keywords: time-stepping; numerical precision; GPU; CUDA; high-order integration.

Full Document Text: PDF version.

Citation Information: BiBTeX database for CSTN Notes.

BiBTeX reference:

@INPROCEEDINGS{CSTN-120,
  author = {K. A. Hawick and D. P. Playne and M. G. B. Johnson},
  title = {Numerical Precision and Benchmarking Very-High-Order Integration
	of Particle Dynamics on GPU Accelerators},
  booktitle = {Proc. International Conference on Computer Design (CDES'11)},
  year = {2011},
  number = {CDE4469},
  pages = {83-89},
  address = {Las Vegas, USA},
  month = {18-21 July},
  publisher = {CSREA},
  institution = {Computer Science, Massey University},
  timestamp = {2011.05.16}
}


[ CSTN Index | CSTN BiBTeX ]