Energy-efficient scheduling on homogeneous multiprocessor platforms
Proceedings of the 2010 ACM Symposium on Applied Computing
Proceedings of the Conference on Design, Automation and Test in Europe
ACM Transactions on Design Automation of Electronic Systems (TODAES)
Scheduling of stream-based real-time applications for heterogeneous systems
Proceedings of the 2011 SIGPLAN/SIGBED conference on Languages, compilers and tools for embedded systems
Platform synthesis and partitioning of real-time tasks for energy efficiency
Journal of Systems Architecture: the EUROMICRO Journal
E-AHRW: An Energy-Efficient Adaptive Hash Scheduler for Stream Processing on Multi-core Servers
Proceedings of the 2011 ACM/IEEE Seventh Symposium on Architectures for Networking and Communications Systems
Energy- and performance-aware scheduling of tasks on parallel and distributed systems
ACM Journal on Emerging Technologies in Computing Systems (JETC)
Throughput-constrained voltage and frequency scaling for real-time heterogeneous multiprocessors
Proceedings of the 28th Annual ACM Symposium on Applied Computing
| Hi-index | 0.00 |
In this paper, we focus on joint energy and performanceoptimization for streaming applications on multiprocessorSystems-on-Chip by combining task-level coarse-grainedsoftware pipelining with DVS (Dynamic Voltage Scaling)and DPM (Dynamic Power Management) techniques withthe considerations of transition overhead, inter-processorcommunication and discrete voltage levels. We proposea two-phase approach to solve the problem. In the firstphase, we propose a coarse-grained task parallelization algorithm called RDAG to transform a periodic dependenttask graph into a set of independent tasks based on the retiming technique[19]. In the second phase, we propose anovel scheduling algorithm called SpringS that works likea spring by iteratively adjusting task scheduling and voltage selection by combining DVS and DPM. We conductexperiments with a set of benchmarks from E3S [10] andTGFF [27]. The experimental results show that our techniquecan achieve 49:8% energy saving on average comparedwith the approach in [20], that applied DVS andDPM without software pipelining. In addition, given a tighttiming constraint, our technique can obtain a feasible solution while the approach in [20] cannot.