A Multipurpose Clustering Algorithm for Task Partitioning in Multicore Reconfigurable Systems

In recent years, multicore systems have become a dominant architecture, introducing new challenges that need to be  addressed in order to take full advantage of their efficiency. Reconfigurable computing has also received a great deal of attention due to its ability to increase the performance of an application through hardware execution, while retaining the flexibility of a software solution. Grouping tasks within an application contributes to coarse-grained partitioning, which can eventually improve the performance of the system. In this paper, we introduce a clustering framework along with a flexible multi-purpose clustering algorithm that initiates task clustering at the functional level based on dynamic profiling information. The clustering framework can be used as the basic step to modify the granularity of tasks in the hardware/software partitioning and scheduling phases. As a result, an elaborate mapping onto the system resources and possibly a higher degree of task parallelism becomes feasible. The framework particularly targets two objectives, 1) to form workload-balanced and 2) loosely-coupled clusters. We evaluated its efficiency using MJPEG as a case study. The experimental results comply with the desired clustering metrics defined through the objectives.