Wednesday, October 12, 2022 11:00 MLIT Conference Hall, 5th floor N. N. Voytishin Mathematical methods and software packages for the reconstruction of charged particle tracks in the BM@N and CMS experiments Abstract: One of the most important problems in experimental high-energy physics is the reconstruction of particle trajectories recorded by tracking detectors of the experimental setups. The quality of the reconstruction directly affects the reliability and accuracy of the subsequent identification of particles and approximation of their momentum. An increase in the luminosity and energy of the beam at the LHC complicates the reconstruction of the trajectories of detected particles. This imposes special requirements on the development of software systems for reconstruction in track detectors of all experiments at the LHC, including the tracking in the Cathode Strip Chambers (CSC) of the CMS experiment. The BM@N experiment is the first stage of the NICA mega-science project. To obtain the first physical results in this experiment, it is necessary to reconstruct the trajectories of charged particles after the analyzing magnet, where the main detectors are the drift chambers (DCH), for which it was necessary to develop the full reconstruction software, from the formation of hits to the complete reconstruction of particle trajectories after the magnet. The report will present the developed reconstruction procedures for the mentioned detector systems and the results of their work for various types of data and acquisition conditions. M. I. Zuev Development of the software and information environment of the "Govorun" Supercomputer for conducting and accelerating applied research carried out at JINR Abstract: The report provides an overview of the work carried out in the framework of the development and support of the software and information environment of the HybriLIT heterogeneous platform, including the installation, configuration, debugging and support of software for computing on the HybriLIT heterogeneous platform (education and testing polygon, “Govorun” supercomputer) both for the NICA experiment and for the tasks of specialists from the JINR Laboratories and Member States. In particular, the results on the creation of a multi-layered system for processing and storing data on the “Govorun” supercomputer, as well as on research on the use of various file systems (Lustre, BeeGFS, DAOS) at the test sites deployed for these purposes, are presented. The results of the following work are also demonstrated: - for a joint project with LRB, an ecosystem for processing, analyzing and visualizing MRI/fMRI data is deployed and tested. Processing data in parallel mode allows one to significantly reduce the computation time. - for a joint project with BLTP on the study of systems based on Josephson junctions, a methodology for developing software modules within the ML/DL/HPC ecosystem based on JupyterHub is presented; it enables not only to carry out calculations, but also to visualize the results of the study and accompany them with the necessary formulas and explanations. - within joint work between MLIT, LRB and the Vinča Nuclear Institute (Belgrade, Serbia), the ML/DL/HPC ecosystem is expanded; it makes possible to effectively perform applied calculations due to the integration of the MATLAB-code into Jupyter Notebook. In addition, information on conducting training courses and master classes on parallel programming technologies, held for a wide range of listeners from JINR, its Member States and Russian universities, will be provided.
