Energy efficient scheduling for hard real-time tasks on dynamic pipeline and voltage scaled processor

This paper addresses an energy optimization problem of pipeline depth and voltage scheduling for hard real-time tasks on a dynamic pipeline and voltage scaled (DPVS) processor, which can adjust its pipeline depth and operating voltage dynamically depending on the workload characteristics under the timing constraints. For a set of periodic tasks, we formulate the energy optimization problem as a linear programming (LP) problem. Then, we propose the energy efficient version of widely used fixed-priority and dynamic-priority scheduling methods, which produce a feasible task schedule with optimal processor energy consumption. This paper also provides a set of experimental results to show the effectiveness of the proposed techniques in reducing energy over the existing methods.