Computational Accelerator Physics Grand Challenge: Summary

The United States is now involved in efforts aimed at developing accelerator-driven technologies for transmutation of radioactive waste, disposal of plutonium, energy production, and production of tritium. These technologies have highly beneficial implications for the nation's environment, future energy needs, and security. Also, next-generation spallation neutron sources based on similar technology will play a major role in materials science and biological science research. Additionally, other types of accelerators such as linear colliders (e.g., the Next Linear Collider (NLC)) and linac-driven coherent light sources will have a significant impact on basic and applied scientific research.

For all of these projects, high-resolution modeling far beyond that which has ever been performed by the accelerator community is required to reduce cost and technological risk, and to improve accelerator efficiency, performance, and reliability. Indeed, such modeling is essential to the success of these efforts. For example, high average power linear accelerators must operate with extremely low beam loss (less than 10 parts per billion per meter) to prevent unacceptably high levels of radioactivity. To ensure that this requirement will be met, it is necessary to perform very high-resolution simulations using on the order of 100 million particles in which the beam propagates through kilometers of complicated accelerating structures. These simulations can only be performed on the most advanced high performance computing (HPC) platforms using software and algorithms targeted to parallel and distributed environments. The calculations require performance of hundreds of GFLOPS to TFLOPS, and core memory requirements of hundreds of GBytes. An equally challenging issue is modeling the fields in electromagnetic structures, especially those having a complex three-dimensional nature, requiring calculations with tens to hundreds of millions of mesh points. Again, such calculations can only be performed in high-performance computing environments.

The primary goal of this Grand Challenge is to develop the next generation of accelerator modeling tools for HPC platforms and to apply them to accelerators for next-generation accelerator applications. The development of portable and reusable software will also enable the solution of other challenging modeling problems such as those associated with future linac-driven coherent light sources and the NLC. The new tools will enable the simulation of problems 3 to 4 orders of magnitude larger than has ever been done before; additionally the use of algorithms and software optimized for the HPC environment will make it possible to obtain results quickly and with very high accuracy. The tools are being developed by a collaboration involving U.S. national laboratories (LANL, SLAC), universities (Stanford, UCLA), and high performance computing and communications research centers (NERSC, ACL).

Back to main page.