Slack reclamation for real-time task scheduling over dynamic voltage scaling multiprocessors

Abstract

In the past decades, a number of research results have been reported for energy-efficient task scheduling over uniprocessor and multiprocessor environments. While researchers have started the exploring of slack reclaiming for tasks during run time, little work has been done for multiprocessor cases. This paper proposes a set of multiprocessor energy-efficient task scheduling algorithms with different task remapping and slack reclaiming schemes, where tasks have the same arrival time and share a common deadline. Tasks are reassigned to processors dynamically, and the slack time is reclaimed to slow down the execution speeds of the remaining tasks for energy efficiency. Extensive simulations were performed to provide insights. The energy consumption could be reduced up to 29% in the experiments, compared to the previous work.