In this paper we present a new search method for partitioning and scheduling a set of periodic tasks on a multiprocessor or distributed architecture. The schedule is fixedpriority driven and task migration is not allowed. The aim of this algorithm is to minimize the number of processors used for scheduling a set of tasks. Moreover, we assume that the number of processors obtained by our method is optimal in respect to the holistic analysis. The paper then compares experimental results from the presented method to the FBB-FFD [9] partitioning algorithm in a multiprocessor context.