Software Energy Optimization of Real Time Preemptive Tasks by Minimizing Cache-Related Preemption Costs

An attempt has been made to optimize the software energy of real time preemptive tasks by minimizing the cache related preemption costs which are primarily incurred due to the inter-task interference in the cache. We have presented an algorithm that outputs an ”optimum” task layout which reduces the overall inter-task interference in cache and thus reduces the preemption costs of each task of a given task set. We have compared the result of our estimated layout with that of the random layout generated for benchmark examples for demonstrating the performance of our algorithm.