We consider the realization of matrix-matrix multiplication and propose a hierarchical algorithm implemented in a task-parallel way using multiprocessor tasks on distributed memory. The algorithm has been designed to minimize the communication overhead while showing large locality of memory references. The task-parallel realization makes the algorithm especially suited for cluster of SMPs since tasks can then be mapped to the different cluster nodes in order to efficiently exploit the cluster

more »

... chitecture. Experiments on current cluster machines show that the resulting execution times are competitive with state-of-the-art methods like PDGEMM .