DAG Scheduling and Analysis on Multi-Core Systems by Modelling Parallelism and Dependency

With ever more complex functionalities being implemented in emerging real-time applications, multi-core systems are demanded for high performance, with directed acyclic graphs (DAG) used to model functional dependencies. For a single DAG task, our previous work presented concurrent provider and consumer (CPC) that captures the node-level dependency parallelism, which two key factors of DAG. Based on CPC, scheduling analysis methods were constructed reduce makespan tighten analytical bound task. However, CPC-based cannot support multi-DAGs as interference between DAGs (i.e., inter-task interference) is not taken into account. To address this limitation, article proposes novel multi-DAG approach specifies number cores can utilise so it does incur interference. This achieved by modelling understanding workload distribution system. By avoiding interference, schedule provides full compatibility be applied each reduces pessimism existing analysis. Experimental results show proposed method achieves an improvement up 80% schedulability against original extends, outperforms 60% tightening