PARALLELIZATION METHOD TO SPEED UP THE TRACK RECONSTRUCTION PROGRAM IN THE SPD NICA EXPERIMENT
Amirkhanova G.A. Mansurova M.Y. Burtebayev N.T. Ososkov G.A. Shomanov A.S.
2023Jagiellonian University
Acta Physica Polonica B, Proceedings Supplement
2023#16Issue 2
Reconstruction of trajectories of charged particles (tracking) is one of the actual problems in experiments in high-energy physics. In the tracking program developed by a team of authors of the Al-Farabi Kazakh National University and JINR to process data registered by detectors located in the magnetic field of the experimental setup SPD planned in a complex of the JINR NICA collider, an algorithm is proposed for sifting out false tracks that arise during neural network tracking. This algorithm is based on a threshold criterion that calculates the quality of the helical line fit to the samples that make up the candidate track recognized by the neural network. In this paper, that continues this research, a method is proposed to significantly speed up the algorithm to weed out false tracks by paralleling it. The results of a comparative analysis of the computational speedup when paralleling them with the condition of preserving the efficiency of track reconstruction are shown.
Text of the article Перейти на текст статьи
Department of Artificial Intelligence & Big Data Al-Farabi Kazakh National University, 71 al-Farabi Ave., Almaty, 050040, Kazakhstan
Institute of Nuclear Physics, Ibragimova str.1, Almaty, 050032, Kazakhstan
Nazarbayev University, Kabanbay batyr str. 53, Nur-Sultan, Kazakhstan
Department of Artificial Intelligence & Big Data Al-Farabi Kazakh National University
Institute of Nuclear Physics
Nazarbayev University
10 лет помогаем публиковать статьи Международный издатель
Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026