By Topic

Coordinated power management of periodic real-time tasks on chip multiprocessors

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Vinay Devadas ; Department of Computer Science, George Mason University, Fairfax, VA 22030, USA ; Hakan Aydin

In this paper, we undertake the problem of minimizing system-level energy on chip-multicore processors (CMPs) executing a periodic real-time workload. Our framework has two components: i.) a static phase that selects a subset of cores upon which the workload can be executed without dissipating excessive static power and performs task-to-core allocation, ii.) a dynamic phase that involves managing the selected cores at run-time through coordinated power management framework that exploits Dynamic Voltage and Frequency Scaling (DVFS) as well as multiple idle states offered by modern CMP architectures, to reduce the dynamic power. We explicitly consider the unique traits of the currently available CMP architectures that distinguish them from multiprocessors, including the unique voltage level shared by the cores and its implications for DVFS. We identify the global energy-efficient frequency which indicates the minimum frequency level at which concurrent execution on multiple cores should take place to preserve the efficiency of DVFS. Then we propose two algorithms CVFS and CVFS* to minimize the dynamic energy consumption through concerted use of DVFS and idle states. Our experimental evaluation indicates that our framework can provide significant gains in system energy.

Published in:

Green Computing Conference, 2010 International

Date of Conference:

15-18 Aug. 2010