SimMatrix: SIMulator for MAny-Task computing execution fabRIc at eXascale

Exascale computing have challenges, most of which can be potentially addressed by Many-task computing paradigm through efficient task execution frameworks that are several orders of magnitude beyond current batch schedulers. This paper proposes a light-weight discrete event simulator, SimMatrix, which simulates distributed job scheduler comprising of millions of nodes and billions of cores/tasks. We validated SimMatrix against MATRIX up to 4K-cores, running on an IBM Blue Gene/P system, and compared SimMatrix with SimGrid and GridSim in terms of resource consumption at scale. Results show that SimMatrix consumes up to two-orders of magnitude lower memory per task, and at least one-order of magnitude (and up to four-orders of magnitude) lower time per task overheads.