A Temporal Dependency Aware Approach for Scheduling Real-Time Tasks on Multi-Core Platforms
Zijun Han, Guangzhi Qu, Бо Лю, Feng Zhang · 2019
The growing demand for computing power in many sectors of industry calls for the utilization of multi-core processors. In order to take advantage of multi-core computing resources, efficient and effective multi-core task scheduling strategies are needed. Directed acyclic graphs (DAG) are commonly used to represent the task execution dependencies that any pair of tasks with dependencies need to be executed in the specified order. In this paper, we investigated and improved the DAGs based task model to consider the temporal property of data dependencies between tasks that the second task does not have to wait until the completion of the precedent task. Based on such a model, we proposed two scheduling strategies: static decomposed scheduling (SDS) and dynamic decomposed scheduling (DDS). The experimental results on both synthesized tasks and real industrial embedded programs show that SDS and DDS achieved good performance when compared with some of the state-of-the-art scheduling strategies.