Basic Design-Time Scheduling

In this chapter, we present the development of a novel task scheduling algorithm for the design-time scheduling phase of our two-phase TCM framework. Scheduling has different meanings in different contexts. In this chapter, scheduling has two meanings. First, assigning every thread node of a given TF to one of the processors of a given platform. Second, deciding the start time to execute every node on its assigned processor. We have two assumptions for this scheduling problem. First, whether a node can be executed on a certain processor should be known. Second, if a node can be executed on a certain processor, the execution time should be known, calculable or predictable. Based on the second assumption, having decided the start time of executing a node on a certain processor, the scheduler can derive the finish time of such an execution immediately. Therefore, the execution order of all the nodes assigned to each processor is known. In summary, scheduling in our context means node-to-processor assignment and ordering.