Investigating the Potential of Energy-savings Using a Fine-grained Task Based Programming Model on Multi-cores

In this paper we study the relation between energy-efficiency and parallel executions when implemented with a fine-grained task-centric programming model. Using a simulation framework comprised of an architectural simulator and a power and area estimation tool, we have investigated the potential energy-savings when employing parallelism on multi-cores system. In our experiments with 2 8 multi-cores systems, we employed frequency and voltage scaling in order to keep the relative performance of the systems constant and measured the energy-efficiency using the Energy-delay-product. Also, we compared the energy consumption of the parallel execution against the serial one. Our results show that through judicious choice of load balancing parameters, significant improvements of around 200 % in energy consumption can be acheived.