No abstract

No abstract

We investigate aspects of the performance of the Eh1EW instance of the Hierarchical PRAM (HPhtAM) model, a recursively pa:rtitionable PRAM, on the 2D mesh architecture via analysis and simulation experiments. Since one of the ideas behind the HPhtAM is to systematically exploit locality in order to negate the need for expensive communication hardware and thus promote cost-eflective scalability, our design decisions are based on minimizing implementation costs. The Peano indexing scheme is used as a simple and natural means of allowing the dynamic, recursive partitioning of the mesh into arbitrarily-sized sub-meshes, as required b y the H-PRAM. We show that for any sub-mesh the ratio of the largest manhai!tan distance between two noc;!es of the sub-mesh to thoit of the square mesh with a:n identical number of processors is at most $/2, demonstrating the locality preserving properties of the Pecino scheme for arbitrary partitions of the mesh. We provide matching analytical and experimental evidence that the routing required to efficiently implem.ent the M-PRAM with this scheme can be implemented checzply and effectively.

No abstract

No abstract