An Experimental Analysis of DAG Scheduling Methods in Hard Real-time Multiprocessor Systems

International audienceThe scheduling of real-time parallel tasks on multiprocessor systems is more complicated than the one of independent sequential tasks, specially for the Directed Acyclic Graph (DAG) parallel model. The complexity is due to the structure of the DAG tasks and the precedence constraints between their subtasks. The trivial DAG scheduling method is to apply directly common real-time scheduling algorithms despite their lack of compatibility with the parallel model. Another scheduling method called the stretching method is summarized as converting each parallel DAG task in the set into a collection of independent sequential threads that are easier to be scheduled. In this paper, we are interested in analyzing global preemptive scheduling of DAGs using both methods by showing that both methods are not comparable in the case of using Deadline Monotonic (DM) and Earliest Deadline First (EDF) scheduling algorithms. Then we use extensive simulations to compare and analyze their performance