A Flexible Tool for Evaluating Scheduling, Synchronization and Partitioning Algorithms on Multiprocessors

• Multi-core platforms seem to be the way towards increasing performance of processors. Single-chip multiprocessors (multi-cores) are today the dominating technology for desktop computing. As the multi-cores are becoming the defacto processors, the need for new scheduling and resource sharing protocols has arisen.There are two major types of scheduling under multiprocessor/multi-core platforms. Global scheduling, under which migration of tasks among processors is allowed, and partitioned scheduling under which tasks are allocated onto processors and task migration is not allowed. The partitioned scheduling protocols suffer from the problem of partitioning tasks among processors/cores, which is a bin-packing problem. Heuristic algorithms have been developed for partitioning a task set on multiprocessor platforms.However, taking such technology to an industrial setting, it needs to be evaluated such that appropriate scheduling, synchronization and partitioning algorithms are selected.In this paper we present our work on a tool for investigation and evaluation of different approaches to scheduling, synchronization and partitioning on multi-core platforms. Our tool allows for comparison of different approaches with respect to a number of parameters such as number of schedulable systems and number of processors required for scheduling.The output of the tool includes a set of information and graphs to facilitate evaluation and comparison of different approaches.

Publikations- och innehållstyp

ref (ämneskategori)

kon (ämneskategori)