"Examination of supercomputer system jobs flow dynamic characteristics"
Antonov A.S., Zhumatii S.A., Nikitenko D.A., Stefanov K.S., Teplov A.M., Shvets P.A.

The article presents the supercomputer jobs flow dynamic characteristics monitoring system implemented on the "SKIF MSU Chebyshev" supercomputer at present. The proposed approach to the analysis allows one to get an adequate estimate of actual properties of the job flow efficiently and technologically simply. It allows one to gain a better understanding of the supercomputer resources utilization, to identify shortcomings in the supercomputer architecture, and to find possible directions of its optimization.

Keywords: supercomputer, performance, efficiency, dynamic characteristics, supercomputer system load, monitoring, resource manager

Antonov A.S., e-mail: asa@parallel.ru;   Zhumatii S.A., e-mail: serg@parallel.ru;   Nikitenko D.A., e-mail: dan@parallel.ru;   Stefanov K.S., e-mail: cstef@parallel.ru;   Teplov A.M.,, e-mail: alex-teplov@yandex.ru;   Shvets P.A., e-mail: shvets.pavel.srcc@gmail.com – Research Computing Center, Lomonosov Moscow State University; Leninskiye Gory, Moscow, 119991, Russia