Efficient parallel algorithms for graph problems

biconnected components connected components EREW graph algorithms linear speedup minimum spanning trees parallel algorithms radix sort spanning trees tree computation

Mathematics Subject Classification ID

Trees (05C05) Connectivity (05C40) Theory of operating systems (68N25) Graph theory (including graph drawing) in computer science (68R10)

Review text

From the abstract (with some additions in parentheses): We present an efficient technique for parallel manipulation of data structures that avoids memory access conflicts. That is, this technique works on the Exclusive Read/Exclusive Write model of computation, which is the weakest shared memory, MIMD machine model. It is used in a new parallel radix sort algorithm that is optimal for keys whose values are over a small range. Using the radix sort and known results for parallel prefix on linked lists (computing all the products of the prefixes of \(a_ 0\circ a_ 1\circ...\circ a_ n\), where \(\circ\) is any associative operation [\textit{C. P. Kruskal}, \textit{L. Rudolph} and \textit{M. Snir}, The power of parallel prefix, IEEE Trans. Comput. 34, 965-968 (1985)]), we develop parallel algorithms that efficiently solve various computations on trees (tree-recursion and tree-traversal) and ``unicycular graphs (connected digraphs in which each node has in-degree one - such a graph is converted into a forest with the same connected compontents). Finally, we develop parallel algorithms for connected components, spanning trees, minimum spanning trees, and other graph problems (biconnected components, and the conversion of a graph from edge-list to adjacency-list representation: the latter is also applied to compute the product of mappings). All of the graph algorithms achieve linear speedup (time inversely proportional to the number of processors) for all but the sparsest graphs (number of edges proportional to number of vertices).

Recommendations

Cites work

Cited in

(34)

This page was built for publication: Efficient parallel algorithms for graph problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1262781)