NASA Logo

NTRS

NTRS - NASA Technical Reports Server

Collections About News Help Login
Back to Results
Array distribution in data-parallel programsWe consider distribution at compile time of the array data in a distributed-memory implementation of a data-parallel program written in a language like Fortran 90. We allow dynamic redistribution of data and define a heuristic algorithmic framework that chooses distribution parameters to minimize an estimate of program completion time. We represent the program as an alignment-distribution graph. We propose a divide-and-conquer algorithm for distribution that initially assigns a common distribution to each node of the graph and successively refines this assignment, taking computation, realignment, and redistribution costs into account. We explain how to estimate the effect of distribution on computation cost and how to choose a candidate set of distributions. We present the results of an implementation of our algorithms on several test problems.
Document ID
19950003593
Acquisition Source
Legacy CDMS
Document Type
Conference Paper
Authors
Chatterjee, Siddhartha (Research Inst. for Advanced Computer Science Moffett Field, CA, United States)
Gilbert, John R. (Research Inst. for Advanced Computer Science Moffett Field, CA, United States)
Schreiber, Robert (Research Inst. for Advanced Computer Science Moffett Field, CA, United States)
Sheffler, Thomas J. (Research Inst. for Advanced Computer Science Moffett Field, CA, United States)
Date Acquired
September 6, 2013
Publication Date
July 1, 1994
Subject Category
Computer Systems
Report/Patent Number
NASA-CR-196376
RIACS-TP-94-09
NAS 1.26:196376
Meeting Information
Meeting: Annual Workshop on Languages and Compilers for Parallelism
Location: Ithaca, NY
Country: United States
Start Date: August 8, 1993
End Date: August 10, 1993
Accession Number
95N10005
Funding Number(s)
CONTRACT_GRANT: NAS2-13721
Distribution Limits
Public
Copyright
Work of the US Gov. Public Use Permitted.

Related Records

There are no records associated with this record.
No Preview Available