Energy-efficient policies for embedded clusters
LCTES '05 Proceedings of the 2005 ACM SIGPLAN/SIGBED conference on Languages, compilers, and tools for embedded systems
Minimizing expected energy in real-time embedded systems
Proceedings of the 5th ACM international conference on Embedded software
IEEE Transactions on Computers
A unified practical approach to stochastic DVS scheduling
EMSOFT '07 Proceedings of the 7th ACM & IEEE international conference on Embedded software
Minimizing expected energy consumption in real-time systems through dynamic voltage scaling
ACM Transactions on Computer Systems (TOCS)
Attaining soft real-time constraint and energy-efficiency in web servers
Proceedings of the 2008 ACM symposium on Applied computing
Stochastic DVS-based dynamic power management for soft real-time systems
Microprocessors & Microsystems
Energy efficient scheduling for real-time systems with mixed workload
EUC'07 Proceedings of the 2007 international conference on Embedded and ubiquitous computing
Energy-efficient tasks scheduling algorithm for real-time multiprocessor embedded systems
The Journal of Supercomputing
Hi-index | 0.00 |
Computing systems, ranging from small battery-operatedembedded systems to more complex general purpose systems,are designed to satisfy various computation demandsin some acceptable time. In doing so, the system is responsiblefor scheduling jobs/requests in a dynamic fashion. Inaddition, with power consumption recently becoming a criticalissue, most systems are also responsible for their ownpower management. In some rare cases, the exact arrivaltime and execution time of jobs/requests is known, leadingto precise scheduling algorithms and power managementschemes. However, more often than not, there is no a-prioriknowledge of the workload. This work evaluates dynamicvoltage scaling (DVS) policies for power management insystems with unpredictable workloads. A clear winner isidentified, a policy that reduces the energy consumption oneorder of magnitude compared to no power management andup to 40% (in real-life traces) and 50% (in synthetic workloads)compared to the second-best evaluated scheme.