Speaker
Description
OPAL is an open-source, parallel particle accelerator simulation framework developed and maintained primarily by the Paul Scherrer Institute (PSI). It provides high-fidelity modelling of charged particle beams, including space-charge effects, beam dynamics, and interactions with accelerator components, and is widely used for the design and analysis of high-intensity accelerator systems. OPAL-X is a next-generation reimplementation of OPAL designed to exploit modern heterogeneous computing platforms, including GPUs as well as conventional CPU-based clusters. However, for large simulations (synchrotrons for example), solving the Poisson equation for space-charge on a full three-dimensional mesh can be computationally prohibitive, motivating the use of a reduced-dimensionality approach that preserves the essential beam physics while significantly lowering the computational cost. To this end, a solution that separates the longitudinal from the transverse solutions has been proposed, referred to as a 2.5D solver. The domain is split into an array of 2D transverse slices in Frenet-Serret coordinates, and the Poisson equation is solved in 2D independently for each of them to find the transverse fields. In the longitudinal direction, the fields depend on the local line density and the shape of the beam pipe (referred to as a λ(z) model, as used in PyHeadTail). Here the results of implementing such an algorithm for OpalX are report