Online 4-Dimensional Reconstruction of Time-Slices in the CBM Experiment

Abstract

The main focus of the CBM experiment (FAIR, Darmstadt, Germany) is the measurement of very rare probes, which require interaction rates of up to 10 MHz. It makes it mandatory to perform the full online event reconstruction at the first level trigger and to operate with huge data rates of up to 1 TB/s. CBM will have a continuous beam without bunch structure, that means collisions may overlap in time, making the traditional event-based approach not applicable. It requires full online event reconstruction and selection to be done in 4D, including time. The standalone First Level Event Selection (FLES) package has been created for the CBM experiment. It contains all reconstruction stages: track finding, track fitting, short-lived particles finding and event selection. For track reconstruction the Cellular Automaton (CA) method is used, that allows to resolve tracks from a time-slice in event-corresponding groups. The algorithm is intrinsically local and the implementation is both vectorized within a core and parallelized between CPU cores. The CA track finder shows a strong scalability on many-core systems. The speed-up factor of 10.6 was achieved on a CPU with 10 physical cores using hyper-threading.