1. Enhancing Plasma Wakefield and E-cloud Simulation Performance Using a Pipelining Algorithm.
- Author
-
Feng, B., Huang, C., Decyk, V., Mori, W. B., Katsouleas, T., and Muggli, P.
- Subjects
- *
PLASMA accelerators , *ELECTRON beams , *SIMULATION methods & models , *DATA pipelining , *ALGORITHMS , *PARALLEL processing - Abstract
Modeling long timescale propagation of beams in plasma wakefield accelerators at the energy frontier and in electron clouds in circular accelerators such as CERN-LHC requires faster and more efficient simulation codes. Simply increasing the number of processors does not scale beyond one-fifth of the number of cells in the decomposition direction. A pipelining algorithm applied on fully parallelized code QuickPIC is suggested to overcome this limit. The pipelining algorithm uses many groups of processors and optimizes the job allocation on the processors in parallel computing. With the new algorithm, it is possible to use on the order of 102 groups of processors, expanding the scale and speed of simulations with QuickPIC by a similar factor. © 2006 American Institute of Physics [ABSTRACT FROM AUTHOR]
- Published
- 2006
- Full Text
- View/download PDF