COMMUNICATION OPERATIONS ON COARSE-GRAINED MESH ARCHITECTURES

In this paper we consider three frequently arising communication opera tions, one-to-all, all-to-one, and all-to-all, We describe architectur e-independent solutions for each operation, as well as solutions tailo red towards the mesh architecture, We show how the relationship among the parameters of a parallel machine and the relationship of these par ameters to the message size determines the best solution. We discuss p erformance and scalability issues of our solutions on the Intel Touchs tone Delta. Our results show that in order to cover a broad range of s calability for a particular operation, multiple solutions should be em ployed.