Abstract

We present a general data parallel formulation for highly irregular problems in High Performance Fortran (HPF). Our formulation consists of(1) a method for linearizing irregular data structures (2) a data parallel implementation (in HPF) of graph partitioning algorithms applied to the linearized data structure, (3) techniques for expressing irregular communication and nonuniform computations associated with the elements of linearized data structures.We demonstrate and evaluate our formulation on a parallel, hierarchical N--body method for the evaluation of potentials and forces of nonuniform particle distributions. Our experimental results demonstrate that efficient data parallel (HPF) implementations of highly nonuniform problems are feasible with the proper language/compiler/runtime support. Our data parallel N--body code provides a much needed "benchmark" code for evaluating and improving HPF compilers.

Highlights

  • Data parallel programming provides an effective way to write maintainable, portable, and scalable parallel codes

  • In this paper we present and evaluate a solution to the disparity between the simple array data structures in data parallel languages and the complicated data structures often found in irregular problems, such as trees and other hierarchical structures

  • In addition to the above heuristics, we developed an extension of Orthogonal recursive bisection (ORB) called rotational recursive bisection (RRB)

Read more

Summary

Introduction

Data parallel programming provides an effective way to write maintainable, portable, and scalable parallel codes. It has enjoyed great success in solving many structured problems, such as dense matrix computations, finite difference calculations, nonadaptive multi–grid [8] and nonadaptive hierarchical methods for the potential and force field evaluation of particle interactions [13, 14]. The data parallel programming paradigm has been used successfully for unstructured finite–element problems [20, 16, 17]. Sloan Research Fellowship, and an Intel research grant

Methods
Results
Conclusion
Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.