TY - GEN
T1 - Evaluating conditional statements in embedded system software
T2 - Proceedings of the International Conference on Embedded Systems and Applications ESA'04 - Proceedings of the International Conference on VLSI, VLSI'04
AU - Vallerio, Keith S.
AU - Jha, Niraj K.
PY - 2004
Y1 - 2004
N2 - Mobile embedded systems occupy an important position in modern life since they allow people to communicate, work, and entertain themselves anywhere at any time. This mobility comes at a price. These systems rely on limited battery supplies, which makes energy consumption an important design constraint. As a result, researchers have spent much time investigating methods to reduce the amount of energy consumed by hardware and have made signinciant progress. Since both hardware and software contribute to system functionality, it is logical that both are good targets for energy reduction. Thus, software energy consumption has recently become an important research area. While much of the previous work focuses on data flow optimizations, the work presented in this paper presents a systematic methodology for optimizing the control flow of control-flow intensive embedded system software. An equation for determining the optimal implementation of multi-way selections is presented along with a technique, called condition implantation, for reducing system energy consumption. Our experiments indicate that processor energy can be reduced by up to 39.7% with an average reduction of 24.2%, whereas the total system energy can be reduced by up to 34.1%, with an average saving of 19.0%. This work is the first to focus on reducing system energy based on control-flow optimization.
AB - Mobile embedded systems occupy an important position in modern life since they allow people to communicate, work, and entertain themselves anywhere at any time. This mobility comes at a price. These systems rely on limited battery supplies, which makes energy consumption an important design constraint. As a result, researchers have spent much time investigating methods to reduce the amount of energy consumed by hardware and have made signinciant progress. Since both hardware and software contribute to system functionality, it is logical that both are good targets for energy reduction. Thus, software energy consumption has recently become an important research area. While much of the previous work focuses on data flow optimizations, the work presented in this paper presents a systematic methodology for optimizing the control flow of control-flow intensive embedded system software. An equation for determining the optimal implementation of multi-way selections is presented along with a technique, called condition implantation, for reducing system energy consumption. Our experiments indicate that processor energy can be reduced by up to 39.7% with an average reduction of 24.2%, whereas the total system energy can be reduced by up to 34.1%, with an average saving of 19.0%. This work is the first to focus on reducing system energy based on control-flow optimization.
UR - http://www.scopus.com/inward/record.url?scp=12744278477&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=12744278477&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:12744278477
SN - 1932415416
SN - 9781932415414
T3 - Proceedings of the International Conference on Embedded Systems and Applications ESA'04 - Proceedings of the International Conference on VLSI, VLSI'04
SP - 63
EP - 69
BT - Proceedings of the International Conference on Embedded Systems and Applications ESA'04 - Proceedings of the INternational Conference on VLSI, VLSI'04
A2 - Arabnia, H.R.
A2 - Guo, M.
A2 - Yang, L.T.
Y2 - 21 June 2004 through 24 June 2004
ER -