MODELING THE POWER CONSUMPTION OF A CLIENT DEVICE MULTICORE PROCESSOR DURING ONLINE INTERCONNECTION

Keywords: modeling the power consumption, CPN Tools, multicore CPU, online video conference.

Abstract

During a pandemic, a large number of users are forced to work not at regular workplaces, but at home, using video conferencing and other online means of intercommunication. In this case, not only desktop computers but also mobile devices are often used. Wherein, for example, the application Zoom can diagnose insufficient connection speed, even if one session is established from one Internet connection. The paper shows that the reason may be the overload of the multicore processor's cluster on the client device. However, the functions of energy saving, embedded in the last actual versions of operation systems, also do not provide complete success. Often, it is needed maximum using the resource of the mobile application, but at the same, time it is necessary to keep up-to-date information about the battery charge. Increasing the number of cores in modern processors does not significantly improve the quality of communication in online communication, because the same type (multimedia) applications are still distributed to the same cluster. In this study, by modeling the power consumption of each core of such a processor, built on the big.LITTLE architecture, the results are achieved, according to which it is possible to adjust the power supply modes of a multi-core processor. The developed software application uses a new power consumption model optimized for modern multi-core processors. Using this model allows you to increase the time of use of the device without recharging to 6.8%, by making timely decisions about CPU load management (e.g., disabling processes and components that overload the CPU or run in the background). The positive difference between the developed software application and the existing schedulers is that this application does not overload client devices and does not affect running processes, i.e., does not slow them down. The power-efficient CPU loading
on the client device will ensure the non-breaking of the communication session during video conferences providing processes in the production industries, the remote provision of educational and medical services.

Author Biographies

Ірина Миколаївна Журавська, Petro Mohyla Black Sea National University

Dr. of Science (Techn.), Professor, Acting Professor of the Department of Computer Engineering

Катерина Олександрівна Обухова, Petro Mohyla Black Sea National University

Master of Computer Science, Professor, Acting Lecturer of the Department of Computer Engineering

Володимир Юрійович Савінов, Petro Mohyla Black Sea National University

y, Ph.D. (Techn.), Senior Lecturer of the Department of Computer Engineering

Published
2021-09-07
How to Cite
Журавська, І., Обухова, К., & Савінов, В. (2021). MODELING THE POWER CONSUMPTION OF A CLIENT DEVICE MULTICORE PROCESSOR DURING ONLINE INTERCONNECTION. Electrotechnic and Computer Systems, (35(111), 73-82. https://doi.org/10.15276/eltecs.35.111.2021.6
Section
Computer Systems, Networks and their Components