Recent trends in real-time systems are towardsmulticores and parallel processes in the form of directed acyclicgraphs. The scheduling aspect of such systems has been workedon and many methods are developed. Nevertheless, the need formore efficient approaches which can use fewer number of coreshas not vanished. Semi-partitioned scheduling of hard real-timeparallel tasks in multicores is studied in this paper. Since thereis no benefit in completing a task much before its deadline, afterscheduling a parallel task, if it is beneficial for other tasks, theexecution of this task is further moved towards its deadline,i.e., stretching, to make room for tasks with closer deadlines.A new concept, prior+, load of tasks is used to rank all tasksof each directed acyclic graph and order them for scheduling.The scheduler is offline and the schedule map is used duringrun time. One benefit of this is the reduction of the schedulingoverhead during run time which helps to safely accept loads.The comparative evaluations show the algorithms performanceis superior to the state of the art ones. It also confirmed thatthe new concept of prior+ load of a task is very effective inscheduling real-time directed acyclic graphs and suggests that itcan as well be useful in scheduling workflows.

Keywords