Unified supercomputer infrastructure based on NIKS launched
Media, 27 January 2022
In January 2022, the first test run of the unified supercomputer infrastructure based on the National Research Computer Network of Russia (NIKS) was completed. The unified supercomputer network includes the “Govorun” supercomputer, an object of the scientific infrastructure of the JINR Member States. As its first assignment, an experiment on mathematical modelling was carried out using the unified supercomputer infrastructure for the tasks of the NICA megascience project.
3,000 Monte Carlo data generation and event reconstruction tasks for the MPD experiment were launched during the six-day test run. This is a small volume. In comparison, only one “Govorun” supercomputer performs 10 times more tasks during the same time in its normal operation mode. As a result, all tasks were completed successfully. About 3 million events were generated and reconstructed. The obtained data have been transported to Dubna for further processing and physical analysis. Based on this data analysis, scientific groups of the MPD experiment expect to achieve promising results for the NICA megaproject.
During the next simulation run at the unified supercomputer infrastructure for MPD, there will be several dozens of millions of generated and reconstructed experimental events. Currently, the unified infrastructure is also not idle. The simulation of atomic electron shell properties of superheavy elements is underway.
It should be reminded that last September, JINR, the Joint Supercomputer Centre of the Russian Academy of Sciences, and Peter the Great St. Petersburg Polytechnic University signed an agreement on uniting their three supercomputers into a single extensible research infrastructure based on the National Research Computer Network of Russia (NIKS). The created infrastructure allows the participants to extend their local computing power, to provide access to big data storage and processing tools and distributed data storages, called datahubs, and also to use each other’s computing powers in cases of peak loads. First of all, such an infrastructure is in demand for the tasks of the NICA megasience project.
Operational demonstration of the unified supercomputer infrastructure for the MPD experiment: https://disk.jinr.ru/index.php/s/swQQ5rM3tGs4nr3