Design and implementation of energy-aware application-specific CPU frequency governors for the heterogeneous distributed computing systems
Michał Karpowicz , Piotr Przemyslaw Arabas , Ewa Niewiadomska-Szynkiewicz
AbstractThis paper deals with the design of application-specific energy-aware CPU frequency scaling mechanisms. The proposed customized CPU controllers may optimize performance of data centers in which diverse tasks are allocated to servers with different characteristics. First, it is demonstrated that server power usage can be accurately estimated based on the measurements of CPU power consumption read from the model specific registers (MSRs). Next, a benchmarking methodology derived from the RFC2544 specification is proposed that allows to identify models of CPU workload dynamics. Finally, it is demonstrated how the identified models can be applied in the design of customized energy-aware controllers that dynamically adjust CPU frequency to the application-specific workload patterns. According to the results of experimental studies the customized controllers may outperform standard general-purpose governors of the Linux kernel both in terms of reachable server performance and power saving capabilities.
|Journal series||Future Generation Computer Systems, ISSN 0167-739X|
|Publication size in sheets||0.65|
|Keywords in English||Green computing; DVFS; Data centers; Optimal control; System identification; Linux|
|project||Development of methodology of control, decision support and production management. Project leader: Zieliński Cezary,
, Phone: 5102, start date 19-05-2015, end date 31-12-2016, 504/02233/1031, Completed
[2015/17/B/ST6/01885] Energy-aware computer system for HPC computing. Project leader: Niewiadomska-Szynkiewicz Ewa, , Phone: 3650, start date 18-02-2016, planned end date 17-02-2019, IA/OPUS9/2015, Implemented
|Score|| = 35.0, 27-03-2017, ArticleFromJournal|
= 40.0, 27-03-2017, ArticleFromJournal
|Publication indicators||: 2016 = 3.997 (2) - 2016=4.787 (5)|
* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.