Scholars@Duke
Journal cover image

A refinement methodology for developing Data-Parallel applications

Publication ,  Conference
Nyland, L; Prins, J; Goldberg, A; Mills, P; Reif, J; Wagner, R
Published in: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
January 1, 1996
Published version (DOI)

Data-parallelism is a relatively well-understood form of parallel computation, yet developing simple applications can involve substantial efforts to express the problem in low-level data-parallel notations. We describe a process of software development for dataparallel applications starting from high-level specifications, generating repeated refinements of designs to match different architectural models and performanee constraints, supporting a development activity with cost-benefit analysis. Primary issues are algorithm choice, correctness and efficiency, followed by data decomposition, load balancing and messagepassing coordination. Development of a data-parallel multitarget tracking application is used as a case study, showing the progression from high to low-level refinements. We conclude by describing tool support for the process.

Duke Scholars

John H. Reif
Author John H. Reif Computer Science

Published In

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

DOI

10.1007/3-540-61626-8_18

EISSN

1611-3349

ISSN

0302-9743

ISBN

9783540616269

Publication Date

January 1, 1996

Volume

1123

Start / End Page

145 / 150

Related Subject Headings

  • Artificial Intelligence & Image Processing
  • 46 Information and computing sciences
 

Citation

APA
Chicago
ICMJE
MLA
NLM
Nyland, L., Prins, J., Goldberg, A., Mills, P., Reif, J., & Wagner, R. (1996). A refinement methodology for developing Data-Parallel applications. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1123, pp. 145–150). https://doi.org/10.1007/3-540-61626-8_18
Nyland, L., J. Prins, A. Goldberg, P. Mills, J. Reif, and R. Wagner. “A refinement methodology for developing Data-Parallel applications.” In Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 1123:145–50, 1996. https://doi.org/10.1007/3-540-61626-8_18.
Nyland L, Prins J, Goldberg A, Mills P, Reif J, Wagner R. A refinement methodology for developing Data-Parallel applications. In: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 1996. p. 145–50.
Nyland, L., et al. “A refinement methodology for developing Data-Parallel applications.” Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1123, 1996, pp. 145–50. Scopus, doi:10.1007/3-540-61626-8_18.
Nyland L, Prins J, Goldberg A, Mills P, Reif J, Wagner R. A refinement methodology for developing Data-Parallel applications. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 1996. p. 145–150.
Journal cover image

Published In

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

DOI

10.1007/3-540-61626-8_18

EISSN

1611-3349

ISSN

0302-9743

ISBN

9783540616269

Publication Date

January 1, 1996

Volume

1123

Start / End Page

145 / 150

Related Subject Headings

  • Artificial Intelligence & Image Processing
  • 46 Information and computing sciences