By Topic

Leakage-aware dynamic scheduling for real-time adaptive applications on multiprocessor systems

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
$31 $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

3 Author(s)
Heng Yu ; Nat. Univ. of Singapore, Singapore, Singapore ; Veeravalli, B. ; Yajun Ha

While performance-adaptable applications are gaining increased popularity on embedded systems (especially multimedia applications), efficient scheduling methods are necessary to explore such feature to achieve the most performance outcome. In addition to conventional scheduling requirements such as real-time and dynamic power, emerging challenges such as leakage power and multiprocessors further complicate the formulation and solution of adaptive application scheduling problems. In this paper, we propose a runtime adaptive application scheduling scheme that efficiently distributes the runtime slack in a task graph, to achieve maximized performance under timing and dynamic/leakage energy constraints. A guided-search heuristics is proposed to select the best-fit frequency levels that maximize the additional program cycles of adaptive tasks. Moreover, we devise a two-stage receiver task selection method that runs efficiently at runtime, in order to quickly find the slack distribution targets. Experiments on synthesized tasks and a JPEG2000 decoder are conducted to justify our approach. Results show that our method achieves at least 25% runtime performance increase compared to contemporary approaches, incurring negligible runtime overhead.

Published in:

Design Automation Conference (DAC), 2010 47th ACM/IEEE

Date of Conference:

13-18 June 2010