AMAC's current activities are devoted mainly to the DOE Office of Science's program, Scientific Discovery through Advanced Computing (SciDAC). The SciDAC accelerator modeling project, "Advanced Computing for 21st Century Accelerator Science and Technology" is run by the Office of High Energy and Nuclear Physics in partnership with the Office of Advanced Scientific Computing Research. The goal of this project is develop a new generation of accelerator modeling tools, implemented on parallel supercomputers, and to apply them to DOE accelerator programs of national and international importance. Through AFRD and AMAC, LBNL is both a co-lead laboratory on the project and is playing a key role in the Beam Dynamics and Advanced Accelerator portions of the project.

The goal of this project is to develop optimal discrete models for infinite-dimensional Hamiltonian systems, and to develop time-domain and eigenmode algorithms and solvers for these models. This work is motivated in part by the great success of symplectic integrators for classical finite-dimensional Hamiltonian systems. These have become indispensable for studying long-term behavior including the dynamics of charged particles in plasmas, the dynamics of stars in gravitational systems, and the dynamics of atoms and molecules in materials. Under this project we will develop new methods that can be applied to infinite dimensional systems governed by equations such as Maxwell's equations, the Schrodinger equation, and the nonlinear Schrodinger equation.

This project will provide a computational foundation for the two key scientific areas critical to the development of the next generation of particle and laser-beam based accelerators: (1) Wave-particle interaction in the full six-dimensional phase-space of multi-particle, multi-mode particle-wave systems; and (2) Coulomb and electromagnetic space-charge dominated transport of high intensity charged particle beams. These will have immediate applications in the studies of laser-plasma acceleration, acceleration and transport of heavy ions for fusion, and simulations of intense bunches in circular accelerators including 3D space charge effects.

The computational approaches used in this project fall into two categories: Particle-In-Cell (PIC) methods and fluid models. In both cases, adaptive mesh refinement (AMR) provides a powerful capability to model complex geometries and to put grid points where they are needed, leading to greatly increased computational efficiency. Under this project, we are developing a capability to model gas jets using the Chombo AMR package, and the ability to model laser-plasma interactions by developing new fluid models and enhancing the XOOPIC PIC code. We are also developing an AMR-based capability to model intense beams in accelerators by combining the WARP code with the Chombo package. In addition, we are developing a capability to model high-order nonlinear effects simultaneously with space-charge effects in circular accelerators.