Where academic tradition
meets the exciting future

You are here: TUCS > PUBLICATIONS > Publication Search > Simulation of Embedded Systems...

Simulation of Embedded Systems for Energy Consumption Estimation

Sébastien Lafond, Simulation of Embedded Systems for Energy Consumption Estimation. TUCS Dissertations 113. Turku Centre for Computer Science, 2009.

Abstract:

Technology developments in semiconductor fabrication along with a rapid expansion of the market for portable devices, such as PDAs and mobile phones, make the energy consumption of embedded systems a major problem. Indeed the need to provide an increasing number of computational intensive applications and at the
same time to maximize the battery life of portable devices can be seen as incompatible trends.

System simulation is a flexible and convenient method for analyzing and exploring the performance of a system or sub-system. At the same time, the increasing use of computational intensive applications strengthens the need to maximize the battery life of portable devices. As a consequence, the simulation of embedded systems for energy consumption estimation is becoming essential in order to study and explore the influence of system design choices on the system energy consumption.

The original publications presented in the second part of this thesis propose several frameworks for evaluating the effects of particular system and software architectures on the system energy consumption. From a software point of view Java and C based applications are studied, and from a hardware perspective systems using general purpose processor and heterogeneous platforms with dedicated hardware accelerators are analyzed. Papers 1 and 2 present a framework for estimating the energy consumption of an embedded Java Virtual Machine and show how an accurate energy consumption model of Java opcodes can be obtained. Paper 3 evaluates the cost-effectiveness of Forward Error Correction algorithms in terms of energy consumption and demonstrates that a substantial energy saving is achievable in a DVB-H receiver when a FEC algorithm is used for file downloading scenarios. Paper 4 and 5 present the simulation of heterogeneous platforms and point out the drawback of different mechanisms used to synchronize a hardware accelerator used as a peripheral device. Paper 6 shows that the use of a multi-bank memory architecture can lead to a 20% static energy saving without any software optimization.

Files:

Full publication in PDF-format

BibTeX entry:

@PHDTHESIS{phdLafond09a,
  title = {Simulation of Embedded Systems for Energy Consumption Estimation},
  author = {Lafond, Sébastien},
  number = {113},
  series = {TUCS Dissertations},
  school = {Turku Centre for Computer Science},
  year = {2009},
  ISBN = {978-952-12-2240-5},
}

Belongs to TUCS Research Unit(s): Embedded Systems Laboratory (ESLAB)

Edit publication