The Community for Technology Leaders
Green Image
Issue No. 01 - January-February (2008 vol. 25)
ISSN: 0740-7475
pp: 52-62
Jorgen Peddersen , University of New South Wales
Sri Parameswaran , University of New South Wales
This article illustrates a low-impact, general method for power and energy management of embedded computing systems. Small counters are added to a system to estimate power and energy consumption of the processor and peripherals in parallel with their execution. This methodology has been applied to an existing 32-bit processor, and a range of benchmarks have demonstrated an average power error of 2%, and an energy estimation error of less than 1.5%. The system adds little impact to the design, with only a 4.9% increase in chip area and a 3% increase in average power consumption. The authors showcase the system by processing a few frames of differing JPEG images, and providing an application that dynamically manages power using the proposed approach.
power estimation, macromodeling, counters, energy aware, low-impact processor, runtime power management

S. Parameswaran and J. Peddersen, "Low-Impact Processor for Dynamic Runtime Power Management," in IEEE Design & Test of Computers, vol. 25, no. , pp. 52-62, 2008.
90 ms
(Ver 3.3 (11022016))