Copyright © 2003 Elsevier Science (USA). All rights reserved.
Constant time parallel sorting: an empirical view
Received 21 November 2001;
revised 15 November 2002.
Available online 28 May 2003.
References and further reading may be available for this article. To view references and further reading you must purchase this article.
Abstract
Consider the following problem: If you want to sort n numbers in k (a constant) rounds then how many comparisons-per-round do you need? This problem has been studied carefully and there exist several algorithms and some lower bounds for it. Many of the algorithms are non-constructive. We have embarked on an empirical study of most of the algorithms in the literature, including the non-constructive ones. This paper is an exposition of what we have found. One of our conclusions is that non-constructive algorithms can be useful.
Article Outline
- 1. Introduction
- 2. Definitions and notation
- 3. Empirical methodology
- 4. Nonconstructive methods
- 5. Constructive methods
- 5.1. Merging and sort
- 5.2. Attempts at the k=2 case
- 5.3. The first constructive subquadratic algorithm for k=2
- 5.4. Two simple constructive algorithms
- 5.5. A constructive algorithm via pseudo-random generators
- 6. A randomized algorithm
- 7. A nonconstructive algorithm for sort(2,n,d)
- 8. A constructive algorithm for sort(2,n,2)
- 9. Lower bounds
- 10. Open problems
- 11. Summary of results
- 11.1. Nonconstructive methods for sorting
- 11.2. Constructive methods for sorting
- 11.3. Limited closure sorting
- 11.4. Lower bounds
- 11.5. Comparisons between different algorithms
- Acknowledgements
- References







E-mail Article
Add to my Quick Links

Cited By in Scopus (0)

/m





