PARADEIS: An object library for parallel sparse array computation

0Citations
Citations of this article
2Readers
Mendeley users who have this article in their library.
Get full text

Abstract

Parallel programming of sparse data structure has been recognized to be hard to debug and to require an excessive amount of development time. The paradeis library (PARAllel Data Environment for Irregular Structure) is an object library that aims at providing a dataparallel approach with object oriented languages to ease the treatment of the sparse data structure for SPMD-like programs on Massively Parallel Processors (MPPs) with distributed memory. This paper focuses on the core of communication layout: the conversion of global addresses to communications, and its relationship with partitioning strategies.

Cite

CITATION STYLE

APA

Delaplace, F., & Remy, D. (1999). PARADEIS: An object library for parallel sparse array computation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1557, pp. 153–162). Springer Verlag. https://doi.org/10.1007/3-540-49164-3_15

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free