Multi-core fixed priority DVS scheduling
ICA3PP'12 Proceedings of the 12th international conference on Algorithms and Architectures for Parallel Processing - Volume Part I
The Journal of Supercomputing
Hi-index | 0.00 |
Dynamic Voltage Scaling (DVS) is an effective technique to reduce the energy consumption of CMOS powered embedded systems through software control. However, applying fixed priority DVS algorithms introduces increased number of preemptions, which, in turn results in extra time delay and energy cost. Effectively reducing the number of preemptions is therefore required. In this paper, we propose to integrate preemption threshold to fixed priority DVS scheduling algorithms to reduce such negative impact. Two preemption-aware algorithms \textit{ccFPPT} and \textit{FPPT-WDA} are studied. Performance evaluations in terms of both energy consumption and the number of preemptions are conducted among different fixed priority DVS algorithms, with or without preemption threshold. The experimental results show that our algorithms with preemption threshold can save up to 60\% number of preemptions and 20\% energy consumption over existing DVS algorithms.