Improvements to the Simbuca trapped charged-particle simulation program

N-body, charged-particle simulations are naturally restricted by the computation time available. The simulation time typically scales as O(N2), since the Coulomb interaction has to be calculated between N particles. The Simbuca simulation package, see Ref. [1], reduces this simulation time by many orders of magnitude by utilizing a graphics card instead of a conventional CPU to calculate the Coulomb interaction. Here, the improvements of Simbuca, will be described. Since its publication the simulation time has improved from around an order of magnitude when simulating a low number of particles (512), to two orders of magnitude when simulating a large number of particles (8192). Further upgrades to Simbuca are summarized and an overview of the applications are given, ranging from many ion clouds to highly charged ions and self-bunching effects in an electrostatic mirror.