Accelerator Modelling and Advanced Computing

Applications

Simulation of beam-beam effects in colliders: Beam-beam modeling is crucial to understanding the beam behavior in high intensity colliders and to developing compensation mechanisms to control beam instabilities and emittance blowup. Beam- beam modeling will also be important for understanding and controlling the behavior of collisions in a future linear collider, where operation in the high-disruption regime means that there can be significant luminosity loss, even from small bunch distortions and other effects. Performing large strong-strong beam-beam simulations is extremely challenging because of the huge amount of particle movement between beam-beam collisions. Under SciDAC, we developed the BeamBeam3D code, a comprehensive 3D parallel PIC code for studying beam-beam effects under a wide range of conditions. It is interesting to note that early in the SciDAC1 Accelerator Science and Technology project we reached a milestone by performing the first-ever million-particle, million-turn "strong-strong" (i.e. self-consistent) beam-beam simulation. With SciDAC-2 now under way, we have far exceeded our early accomplishments in beam-beam modeling. Working with staff in the Performance Enhancement and Research Center (PERC) of SciDAC-1, we achieved 100M particle simulations on 1024 processors on the Seaborg computer at NERSC. Besides being able to perform large simulations, we have now added several important new physics capabilities to BeamBeam3D, including multi-bunch, multi-interaction point, and multi-slice modeling. The BeamBeam3D code has been applied to all the colliders of the HEP and NP programs, the Tevatron, PEP-II, and RHIC, as well as the soon-to-be-operating LHC. BeamBeam3D has also been installed at Fermilab. It is being used and enhanced by researchers there to support Fermilab projects, and was also used to compare with experimental results from the VEPP-2M collider.

Simulation of high brightness electron beams and light sources: High brightness electron beams are important to a number of projects across DOE/SC including HEP (linear collider), NP (CEBAF and proposed electron-ion colliders), and the light sources of the BES program. The IMPACT suite of codes developed by AMAC staff members has been applied to several photoinjector projects nationwide including photoinjectors at ANL, BNL, Cornell, FNAL/NIU, JLab, LBNL, and SLAC/LCLS. Our predictive capability has benefited from our ability to run large, 3D simulations. In 2007 we performed the first-ever 1 billion particle simulation of a linac for a light source to accurately study the microwave instability. But, beyond the ability to perform simulations with large numbers of particles and grid points, the development of multi-physics modeling capabilities, and the development of new computational algorithms, have also been crucial. For example, Coulomb collisions are important in certain situations, such as emission from a nano-needle cathode. In some situations, it is essential to be able to model beams with large energy spread; a technique for doing this was developed and implemented in IMPACT under SciDAC-1. And, in certain situations, the beams have a very large geometrical aspect ratio. Under SciDAC-1 AMAC staff members developed a new class of 3D parallel Poisson solvers specially designed to efficiently model beams that have high aspect ratios, work that has continued in collaboration with researchers from Tech-X Corporation.