Battery aware task scheduling for a system-on-a-chip using voltage/clock scaling

Battery lifetime is a critical parameter in the operation of mobile computing devices. The lifetime of such devices is directly dependent on the battery discharge profile. In this paper we address the problem of task scheduling in single processor and multiprocessor systems such that the battery lifetime is maximized. We propose a procedure that achieves this by shaping the current load profile. The shaping algorithm makes extensive use of voltage/clock scaling and is guided by heuristics that are derived from the properties of the battery model. Simulations show that the proposed algorithm improves the battery lifetime significantly.