Optimized Calculation of Timing for Parallel Beam Operation at the FAIR Accelerator Complex

For the new FAIR accelerator complex at GSI the settings management system LSA is used. It is developed in collaboration with CERN and until now it is executed strictly serial. Nowadays the performance gain of single core processors have nearly stagnated and multicore processors dominate the market. This evolution forces software projects to make use of the parallel hardware to increase their performance. In this thesis LSA is analyzed and parallelized using different parallelization patterns like task and loop parallelization. The most common case of user interaction is to change specific settings so that the accelerator performs at its best. For each changed setting, LSA needs to calculate all child settings of the parameter hierarchy. To maximize the speedup of the calculations, they are also optimized sequentially. The used data structures and algorithms are reviewed to ensure minimal resource usage and maximal compatibility with parallel execution. The overall goal of this thesis is to speed up the calculations so that the results can be shown in an user interface with nearly no noticeable latency.