CEAT: A Cluster based Energy Aware Scheduler for Real-Time Heterogeneous Systems
Sanjay Moulik, Zinea Das, Gitimoni Saikia
Abstract
Modern real-time systems based on heterogeneous multicore platforms can efficiently meet the disparate and high computation needs of the applications. The management of energy has become a topic of an incredible enthusiasm for analysts and practitioners during the past few years. Hence, this research presents a heuristic strategy named, CEAT, for energy-aware scheduling of a set of real-time periodic tasks on a DVFS enabled heterogeneous multicore platform. The presented strategy operates in three stages, namely Deadline Partitioning, Task-to-Core Allocation, and Energy-Aware Scheduling. Our experimental analysis shows that CEAT is not only able to successfully schedule more task sets (as high as 3.44% and 21.83%) compared to state-of-the-art [1] but also improve energy savings in the system.