Combining Genetic Algorithms Based Task Mapping and Optimal Voltage Selection for Energy-Efficient Distributed System Synthesis

Energy consumption has become one of dominant design concerns in distributed embedded systems. Dynamic voltage scaling (DVS) techniques have been proved to be quite effective in reducing both dynamic and static power consumption simultaneously. Thus, integrating processing elements (PEs) supporting DVS techniques into distributed embedded systems synthesis becomes an imperative way to make an energy-efficient design budget. In this paper, we propose a synthesis framework, which integrates task mapping, scheduling and voltage selection together to minimize energy consumption of distributed systems. Genetic algorithms based task mapping are deployed to do task mapping in the outer loop, which can maximize the energy saving possibilities. A nonlinear programming based optimal voltage selection engine is then used in the inner loop. Unlike previous approaches, this combination is the first time to be implemented, and relevant experiment results show that, comparing to separate synthesis flows, up to 40% more energy savings are achieved

[1]  D. Chen,et al.  Task scheduling and voltage selection for energy minimization , 2002, Proceedings 2002 Design Automation Conference (IEEE Cat. No.02CH37324).

[2]  Frank Mueller,et al.  Feedback EDF scheduling exploiting dynamic voltage scaling , 2004, Proceedings. RTAS 2004. 10th IEEE Real-Time and Embedded Technology and Applications Symposium, 2004..

[3]  Miodrag Potkonjak,et al.  On-line scheduling of hard real-time tasks on variable voltage processor , 1998, ICCAD.

[4]  Michael Pinedo,et al.  Scheduling: Theory, Algorithms, and Systems , 1994 .

[5]  Petru Eles,et al.  Overhead-conscious voltage selection for dynamic and leakage energy reduction of time-constrained systems , 2004, Proceedings Design, Automation and Test in Europe Conference and Exhibition.

[6]  David E. Goldberg,et al.  Genetic Algorithms in Search Optimization and Machine Learning , 1988 .

[7]  F. Frances Yao,et al.  A scheduling model for reduced CPU energy , 1995, Proceedings of IEEE 36th Annual Foundations of Computer Science.

[8]  Niraj K. Jha,et al.  Power-conscious joint scheduling of periodic task graphs and aperiodic tasks in distributed real-time embedded systems , 2000, IEEE/ACM International Conference on Computer Aided Design. ICCAD - 2000. IEEE/ACM Digest of Technical Papers (Cat. No.00CH37140).

[9]  Saibal Mukhopadhyay,et al.  Leakage current mechanisms and leakage reduction techniques in deep-submicrometer CMOS circuits , 2003, Proc. IEEE.

[10]  M. Hussein,et al.  A 130 nm generation logic technology featuring 70 nm transistors, dual Vt transistors and 6 layers of Cu interconnects , 2000, International Electron Devices Meeting 2000. Technical Digest. IEDM (Cat. No.00CH37138).

[11]  Bashir M. Al-Hashimi,et al.  Considering power variations of DVS processing elements for energy minimisation in distributed systems , 2001, International Symposium on System Synthesis (IEEE Cat. No.01EX526).

[12]  Brian W. Kernighan,et al.  AMPL: A Modeling Language for Mathematical Programming , 1993 .

[13]  Jorge Nocedal,et al.  An Interior Point Algorithm for Large-Scale Nonlinear Programming , 1999, SIAM J. Optim..

[14]  Wayne H. Wolf,et al.  TGFF: task graphs for free , 1998, Proceedings of the Sixth International Workshop on Hardware/Software Codesign. (CODES/CASHE'98).