Parallel tree slicing

Abstract: "Let T be an ordered rooted tree of height h with n vertices. Vertex v is said to be at level i of T if the path from the root to v contains i edges. We introduce the tree slicing problem of parallel computing: compute, for all i [element of] [0, ..., h], the ordered list of vertices...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Katajainen, Jyrki (VerfasserIn), Thorup, Mikkel 1973- (VerfasserIn), Träff, Jesper Larsson (VerfasserIn)
Format: Buch
Sprache:English
Veröffentlicht: København 1995
Schriftenreihe:Datalogisk Institut <København>: DIKU-Rapport 1995,13
Schlagworte:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Abstract: "Let T be an ordered rooted tree of height h with n vertices. Vertex v is said to be at level i of T if the path from the root to v contains i edges. We introduce the tree slicing problem of parallel computing: compute, for all i [element of] [0, ..., h], the ordered list of vertices at level i of T. Tree slicing is a natural generalization of list ranking to trees, and is intended for scheduling level-wise computations on trees, as for example counting the number of vertices at each level. A work-, time-, and space-optimal parallel algorithm is presented that solves the tree slicing problem in O(log n) time and O(n) space with O(n/log n) processors on the EREW PRAM model. The result is obtained by transforming the tree slicing problem to a special integer sorting problem that is solved within these resource bounds."
Beschreibung:5 S.