DocumentCode :
3204817
Title :
Efficient parallel algorithms for selection and searching on sorted matrices
Author :
Sarnath, R. ; He, Xin
Author_Institution :
Dept. of Comput. Sci., State Univ. of New York, Buffalo, NY, USA
fYear :
1992
fDate :
23-26 Mar 1992
Firstpage :
108
Lastpage :
111
Abstract :
Parallel algorithms for more general versions of the well known selection and searching problems are formulated. The authors look at these problems when the set of elements can be represented as an n ×n matrix with sorted rows and columns. The selection algorithm takes O(lognloglogn log* n) time with O(n/log nlog* n) processors on an EREW PRAM. The searching algorithm takes O(loglogn) time with O(n/loglogn ) processors on a CREW PRAM, which is optimal. The authors also show that no algorithm using at most n logc n processors, c⩾1, can solve the matrix search problem in time faster than Ω(log log n)
Keywords :
computational complexity; matrix algebra; parallel algorithms; EREW PRAM; O(loglogn) time; O(lognloglogn log* n) time; O(n/log nlog* n) processors; O(n/loglogn) processors; matrix search problem; parallel algorithms; searching; selection; sorted matrices; space complexity; time complexity; Computational modeling; Computer networks; Computer science; Concurrent computing; Helium; Parallel algorithms; Phase change random access memory; Search problems; Statistics;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Parallel Processing Symposium, 1992. Proceedings., Sixth International
Conference_Location :
Beverly Hills, CA
Print_ISBN :
0-8186-2672-0
Type :
conf
DOI :
10.1109/IPPS.1992.223063
Filename :
223063
Link To Document :
بازگشت