Author: Arumugam, A.
Paper Title Page
THPAB085 Simulations of Coherent Synchrotron Radiation on Parallel Hybrid GPU/CPU Platform 3915
 
  • B. Terzić, D. Duffin, A.L. Godunov
    ODU, Norfolk, Virginia, USA
  • A. Arumugam, T. Islam, D. Ranjan, S. Sangam, M. Zubair
    ODU CS, Norfolk, Virginia, USA
 
  Funding: National Science Foundation 1535641
Coherent synchrotron radiation (CSR) is an effect of self-interaction of an electron bunch as it traverses a curved path. It can cause a significant emittance degradation, as well as fragmentation and microbunching. Numerical simulations of the 2D/3D CSR effects have been extremely challenging due to computational bottlenecks associated with calculating retarded potentials via integrating over the history of the bunch. We present a new high-performance 2D, particle-in-cell code which uses massively parallel multicore GPU/GPU platforms to alleviate computational bottlenecks. The code formulates the CSR problem from first principles by using the retarded scalar and vector potentials to compute the self-interaction fields. The speedup due to the parallel implementation on GPU/CPU platforms exceeds three oders of magnitude, thereby bringing a previously intractable problem within reach. The accuracy of the code is verified against analytic 1D solutions (rigid bunch) and semi-analytic 2D solutions for the chirped bunch. Finally, we use the new code in conjunction with a genetic algorithm to optimize the design of a fiducial chicane.
 
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2017-THPAB085  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPAB086 Long-Term Simulations of Beam-Beam Dynamics on GPUs 3918
 
  • B. Terzić, C.M. Cotnoir, A.L. Godunov, T. Satogata, M. Stefani
    ODU, Norfolk, Virginia, USA
  • A. Arumugam, R.T. Majeti, D. Ranjan, M. Zubair
    ODU CS, Norfolk, Virginia, USA
  • F. Lin, V.S. Morozov, E.W. Nissen, Y. Roblin, T. Satogata, H. Zhang
    JLab, Newport News, Virginia, USA
 
  Funding: Jefferson Lab
Future machines such as the electron-ion colliders (JLEIC), linac-ring machines (eRHIC) or LHeC are particularly sensitive to beam-beam effects. This is the limiting factor for long-term stability and high luminosity reach. The complexity of the non-linear dynamics makes it challenging to perform such simulations which require millions of turns. Until recently, most of the methods used linear approximations and/or tracking for a limited number of turns. We have developed a framework which exploits a massively parallel Graphical Processing Units (GPU) architecture to allow for tracking millions of turns in a sympletic way up to an arbitrary order and colliding them at each turn. The code is called GHOST for GPU-accelerated High-Order Symplectic Tracking. As of now, there is no other code in existence that can accurately model the single-particle non-linear dynamics and the beam-beam effect at the same time for a large enough number of turns required to verify the long-term stability of a collider. Our approach relies on a matrix-based arbitrary-order symplectic particle tracking for beam transport and the Bassetti-Erskine approximation for the beam-beam interaction.
 
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2017-THPAB086  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)