Adaptive energy minimization of embedded heterogeneous system using regression-based learning


Yang, Sheng, Shafik, Rishad Ahmed, Merrett, Geoff V., Stott, Edward, Levine, Joshua, Davis, James and Al-Hashimi, Bashir (2015) Adaptive energy minimization of embedded heterogeneous system using regression-based learning At 25th International Workshop on Power and Timing Modeling, Optimization and Simulation (PATMOS 2015), Brazil. 01 - 04 Sep 2015. 8 pp.

Download

[img] PDF sigproc.pdf - Accepted Manuscript
Download (2MB)

Description/Abstract

Modern embedded systems consist of heterogeneous computing resources with diverse energy and performance trade-offs. This is because the computing resources exercise the application tasks differently, generating varying workloads and energy consumption. As a result, minimizing energy consumption in these systems is challenging as it requires continuous adaptation of application task mapping (i.e. allocating tasks among the computing resources) and dynamic voltage/frequency scaling (DVFS). Existing approaches lack such adaptation with practical validation (Table I).
This paper proposes a novel adaptive energy minimization ap- proach for embedded heterogeneous systems. Fundamental to this approach is a runtime model, generated through regression-based learning of energy/performance trade-offs between different comput- ing resources in the system. Using this model, an application task is suitably mapped on a computing resource during runtime, ensuring minimum energy consumption for a given application performance requirement. Such mapping is also coupled with a DVFS control to adapt to performance and workload variations. The proposed approach is designed, engineered and validated on a Zynq-ZC702 platform, consisting of CPU, DSP and FPGA cores. Using several image processing applications as case studies, our proposed approach can achieve significant energy savings (70% in some cases, i.e. from 43mJ per frame to 13 mJ per frame), when compared to existing approaches.

Item Type: Conference or Workshop Item (Paper)
Venue - Dates: 25th International Workshop on Power and Timing Modeling, Optimization and Simulation (PATMOS 2015), Brazil, 2015-09-01 - 2015-09-04
Related URLs:
Keywords: energy efficiency, dynamic voltage/frequency scaling, runtime optimization, linear regression
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Organisations: Electronics & Computer Science
ePrint ID: 379535
Date :
Date Event
22 July 2015Accepted/In Press
Date Deposited: 05 Aug 2015 13:39
Last Modified: 17 Apr 2017 05:35
Further Information:Google Scholar
URI: http://eprints.soton.ac.uk/id/eprint/379535

Actions (login required)

View Item View Item