An Energy-Efficient Task Scheduler for Multi-core Platforms with Per-core DVFS Based on Task Characteristics

Abstract

Energy-efficient task scheduling is a fundamental issue in many application domains, such as energy conservation for mobile devices and the operation of green computing data centers. Modern processors support dynamic voltage and frequency scaling (DVFS) on a per-core basis, i.e., the CPU can adjust the voltage or frequency of each core. As a result, the core in a processor may have different computing power and energy consumption. To conserve energy in multi-core platforms, we propose task scheduling algorithms that leverage per-core DVFS and achieve a balance between performance and energy consumption. We consider two task execution modes: the batch mode, which runs jobs in batches, and the online mode in which jobs with different time constraints, arrival times, and computation workloads co-exist in the system. For tasks executed in the batch mode, we propose an algorithm that finds the optimal scheduling policy, and for the online mode, we present a heuristic algorithm that determines the execution order and processing speed of tasks in an online fashion. The heuristic ensures that the total cost is minimal for every time interval during a task's execution.