Energy-efficient real-time scheduling of multimedia tasks on multi-core processors.

ABSTRACTIn recent years, various multi-core architectures have become popular selections for the designs of mobile platforms. With the strong computing demands from many multimedia applications, how to energy-efficiently utilize the computing power of mobile platforms without violations of timing constraints has become a critical design problem. In this paper, a data-partitioning-based approach is proposed to explore the parallelism of multimedia workload processing over multiple cores. Dynamic voltage scaling and dynamic power management strategies are both considered in the dynamic scaling of the computing power of cores and the adjustment of the set of active cores, respectively. The practicability and the energy efficiency of the proposed algorithms were evaluated by a series of experiments and simulations, for which we have encouraging results.