A General Data Layout for Distributed Consistency in Data Parallel Applications

Distributed recursive sets: programmability and effectiveness for data intensive applications

“…Thus, starting with the sets at the highest level of recursion, the algorithm maps the data at the next level to increase locality of reference and thus, reduce communication. The main steps of the algorithm proceed as follows Given the partitioning and mapping of data to processors, we use the algorithm described in [3] to maintain data consistency.…”