München 2019 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

HK: Fachverband Physik der Hadronen und Kerne

HK 35: Heavy-Ion Collisions and QCD Phases VII

HK 35.3: Vortrag

Mittwoch, 20. März 2019, 14:45–15:00, HS 12

Application and improvement of Cellular Automaton track finder in the TPC detector — •Grigory Kozlov^{1, 2} and Ivan Kisel^{1, 3} for the CBM collaboration — ¹FIAS, Frankfurt am Main, Germany — ²JINR, Dubna, Russia — ³GSI, Darmstadt, Germany

Track finding procedure is one of the most important part of event reconstruction in high energy physics experiments. Tracking algorithms combine hits into tracks and reconstruct trajectories of particles flying through the detector. Due to the high combinatorics, they are usually considered as the most time consuming tasks. Calculation speed is critical in heavy ion experiments, especially for online reconstruction. Thus, tracking procedure mast be extremely fast, keeping high efficiency at the same time. The Cellular Automaton (CA) algorithm provides a perfect solution for this task. Because of its properties, it provides required speed, efficiency and can be massively parallelised on the modern many core computing platforms. The CA track finder algorithm was investigated in application to the TPC CA track finder in the STAR experiment within the FAIR Phase 0 as a part of preparation to the Beam Energy Scan II (BES II) program. The initial implementation of the TPC CA algorithm is dominated by the data copying, that complicates an efficient utilisation of the advantages of the SIMD unit. In this work we consider several improvements of the CA algorithm, which allows us to get the maximum benefit from using vectorized calculations. New approaches are compatible and interchangeable with existing CA methods. This allows full use of the SIMD-registers and sufficiently high efficiency at the maximal calculation speed.