Energy minimisation techniques for distributed embedded systems

Abstract

It is likely that the demand for embedded computing systems with low energy dissipation will continue to increase. This thesis is concerned with the development and validation of techniques that allow an effective automated design of energy-efficient embedded systems. Special emphasis is placed upon system-level co-synthesis techniques for systems that contain dynamic voltage scalable processors which can trade off between performance and power consumption during run-time.

The first part of the thesis addresses energy minimisation of distributed embedded systems through dynamic voltage scaling (DVS). A new voltage selection technique for single-mode systems based on a novel energy-gradient scaling strategy is presented. This technique, which overcomes limitations of previous voltage scaling approaches, exploits system idle and slack time to reduce the power consumption, taking into account the individual task power dissipation. Numerous benchmark experiments validate the quality of the proposed technique in terms of energy reduction and computational complexity.

The second part of the thesis focuses on the development of genetic algorithm-based co-synthesis techniques (mapping and scheduling) for single-mode systems that have been specifically developed for an effective utilisation of the voltage scaling approach introduced in the first part. The schedule optimisation improves the execution order of system activities not only towards performance, but also towards a high exploitation of voltage scaling to achieve energy savings. The mapping optimisation targets the distribu­tion of system activities across the system components to further improve the utilisation of DVS, while satisfying hardware area constraints. Extensive experiments including a real-life optical flow detection algorithm are conducted, and it is shown that the proposed co-synthesis techniques can lead to high energy savings with moderate computational overhead.

The third part of this thesis concentrates on energy minimisation of emerging distributed embedded systems that accommodate several different applications within a single device, i.e., multi-mode embedded systems. A new co-synthesis technique for multi-mode embedded systems based on a novel operational-mode-state-machine specification is presented. The technique increases significantly the energy savings by considering the mode execution probabilities that yields better resource sharing opportunities.

The new co-synthesis and voltage scaling techniques have been incorporated into the prototype co-synthesis tool LOPOCOS (Low Power Co-Synthesis). The capability of LOPOCOS in efficiently exploring the architectural design space is demonstrated through a system-level design of a realistic smart phone example that integrates a GSM cellular phone transcoder, an MP3 decoder, as well as a JPEG image encoder and decoder.

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata