Date: 09 Jun 2005

Data redistribution in an automatic data distribution tool

* Final gross prices may vary according to local VAT.

Get Access

Abstract

Data distribution is one of the key aspects to consider in a parallelizing environment for Massive Parallel Processors. Automatic data distribution proposals may be categorized as either static or dynamic, depending on whether the distribution of the arrays is allowed to change throughout the execution of the program. This paper describes the features and implementation of the intra-procedural data remapping module implemented in our automatic data distribution research tool. The solution to the remapping problem for a sequence of computational phases consists in selecting a mapping among the possible candidates for each phase and introduce remapping actions between consecutive phases. Control flow information is used to identify how phases are sequenced during the execution of the application.

This research was partially supported by Convex Computer Corporation, CONVEX Supercomputers S.A.E, CEPBA (European Center for Parallelism of Barcelona) and by the Ministry of Education of Spain under contracts TIC-880/92 and TIC-429/95.