List ranking on processor arrays

List ranking finds for each cell in a linked list the number of cells that precede it in the list. This paper presents a work-efficient list-ranking algorithm for fine-grained processor arrays. This algorithm runs on an array of n/log 2 n processors with the expected run-time of O(log 2 n). As list...

Full description

Saved in:
Bibliographic Details
Published inThe Journal of systems and software Vol. 55; no. 2; pp. 185 - 192
Main Author Hasan, Cam
Format Journal Article
LanguageEnglish
Published New York Elsevier Inc 29.12.2000
Elsevier Sequoia S.A
Subjects
Algorithms
Arrays
List ranking
Parallel algorithm
Processor array
Studies
Work-efficient PRAM algorithm
List ranking
Parallel algorithm
Processor array
Work-efficient PRAM algorithm
Online AccessGet full text
ISSN0164-1212
1873-1228
DOI10.1016/S0164-1212(00)00069-8

Cover

More Information
Summary:List ranking finds for each cell in a linked list the number of cells that precede it in the list. This paper presents a work-efficient list-ranking algorithm for fine-grained processor arrays. This algorithm runs on an array of n/log 2 n processors with the expected run-time of O(log 2 n). As list ranking is highly communication intensive, the proposed algorithm is able to reduce communication cost among processors by assigning sublists, instead of arbitrary cells, of a linked list to each processor. The proposed algorithm is also capable of keeping all processors busy during the whole list-ranking process in order to utilize all processors efficiently.
Bibliography:SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 14
ObjectType-Article-2
content type line 23
ISSN:0164-1212
1873-1228
DOI:10.1016/S0164-1212(00)00069-8