| Paper |
Title |
Page |
| THPAS019 |
A Beam Dynamics Application Based on the Common Component Architecture
|
3552 |
| |
- D. R. Dechow
- D. T. Abell, P. Stoltz
Tech-X, Boulder, Colorado
- J. F. Amundson
Fermilab, Batavia, Illinois
- L. Curfman McInnes, B. Norris
ANL, Argonne, Illinois
|
|
| |
Funding: Department of Engergy, Office of Advanced Scientific Computing Research, SBIR grant: DE-FG02-06ER84520
A component-based beam dynamics application for modeling collective effects in particle accelerators has been developed. The Common Component Architecture (CCA) software infrastructure was used to compose a new Python-steered accelerator simulation from a set of services provided by two separate beam dynamics packages (Synergia and MaryLie/Impact) and two high-performance computer science packages (PETSc and FFTW). The development of the proof-of-concept application was accomplished via the following tasks: - addressing multilanguage interoperability in the MaryLie/Impact code with Babel;
- creating components by making the selected software objects adhere to the Common Component Architecture protocol;
- assemblying the components with a newly developed, Component Builder gui; and
- characterizing the performance of the space charge (Poisson) solver that was originally used in Synergia 1.0 versus the PETSc-based space charge solver that has been developed for Synergia2.
The resulting beam dynamics application will allow the Synergia2 framework to evolve simultaneously with the modeling and simulation requirements of the International Linear Collider.
|
|
| TUZBC02 |
SciDAC Frameworks and Solvers for Multi-physics Beam Dynamics Simulations
|
894 |
| |
- J. F. Amundson
- D. R. Dechow
Tech-X, Boulder, Colorado
- J. Qiang, R. D. Ryne
LBNL, Berkeley, California
- P. Spentzouris
Fermilab, Batavia, Illinois
|
|
| |
The need for realistic accelerator simulations is greater than ever before due to the needs of design projects such as the ILC and optimization for existing machines. Sophisticated codes utilizing large-scale parallel computing have been developed to study collective beam effects such as space charge, electron cloud, beam-beam, etc. We will describe recent advances in the solvers for these effects and plans for enhancing them in the future. To date the codes have typically applied to a single collective effect and included just enough of the single-particle dynamics to support the collective effect at hand. We describe how we are developing a framework for realistic multi-physics simulations, i.e., simulations including the state-of-the-art calculations of all relevant physical processes.
|
|
|
Slides
|
|