Concurrent threads and optimal parallel minimum spanning trees algorithm
- 1 March 2001
- journal article
- Published by Association for Computing Machinery (ACM) in Journal of the ACM
- Vol. 48 (2) , 297-323
- https://doi.org/10.1145/375827.375847
Abstract
This paper resolves a long-standing open problem on whether the concurrent write capability of parallel random access machine (PRAM) is essential for solving fundamental graph problems like connected components and minimum spanning trees in O (log n ) time. Specifically, we present a new algorithm to solve these problems in O (log n ) time using a linear number of processors on the exclusive-read exclusive-write PRAM. The logarithmic time bound is actually optimal since it is well known that even computing the “OR” of n bit requires Ω(log n time on the exclusive-write PRAM. The efficiency achieved by the new algorithm is based on a new schedule which can exploit a high degree of parallelism.Keywords
This publication has 22 references indexed in Scilit:
- A minimum spanning tree algorithm with inverse-Ackermann type complexityJournal of the ACM, 2000
- A randomized linear-time algorithm to find minimum spanning treesJournal of the ACM, 1995
- A bridging model for parallel computationCommunications of the ACM, 1990
- An efficient and fast parallel-connected component algorithmJournal of the ACM, 1990
- Fibonacci heaps and their uses in improved network optimization algorithmsJournal of the ACM, 1987
- Upper and Lower Time Bounds for Parallel Random Access Machines without Simultaneous WritesSIAM Journal on Computing, 1986
- An Efficient Parallel Biconnectivity AlgorithmSIAM Journal on Computing, 1985
- On efficient parallel strong orientationInformation Processing Letters, 1985
- Efficient parallel algorithms for some graph problemsCommunications of the ACM, 1982
- Computing connected components on parallel computersCommunications of the ACM, 1979