DAG Scheduling for Clustered Many-Core Processor Considering Execution-Time-Reduction Effectiveness

In recent years, there has been a surge in research on autonomous systems such as unmanned aerial vehicles and autonomous-driving systems. These systems require high computing power with low power consumption and strict real-time requirements. Correspondingly, clustered many-core processors are attracting attention as a solution to these requirements. In this study, we propose a core allocation method for DAG scheduling using clustered many-core processors. Specifically, we propose two methods for selecting nodes to be allocated based on the execution-time-reduction effectiveness, assuming that each node in the DAG can perform intra-node processing in parallel. In addition, we compared the response times obtained from scheduling using each of the two methods to determine which method is superior.