
Accession Number : ADA113104
Title : Implementation of the GibbsPooleStockmeyer Algorithm and the GibbsKing Algorithm.
Descriptive Note : Technical rept.,
Corporate Author : BOEING COMPUTER SERVICES CO TUKWILA WA
Personal Author(s) : Lewis,John Gregg
PDF Url : ADA113104
Report Date : Jan 1982
Pagination or Media Count : 20
Abstract : Implementations of two effective matrix reordering algorithms were published in this paper as Algorithms 508, which implements the GibbsPooleStockmeyer algorithm for reducing the bandwidth of a matrix, and 509, which implements the GibbsKing algorithm for reducing the profile of a matrix. Reduction of matrix profile is more often required than bandwidth reduction, but Algorithm 509 is substantially slower than Algorithm 508. Consequently, the GibbsPooleStockmeyer algorithm has been recommended and used in contexts where the better profile reduction provided by the GibbsKing algorithm would be more appropriate. In addition, Algorithms 508 and 509 both contain unnecessary restrictions on problem size and provide little error checking. The authors describe a new FORTRAN implementation of both reordering algorithms which is portable, faster, more reliable and uses less storage than the original implementations. The new implementation of the GibbsKing algorithm is much faster than Algorithm 509, generally slightly faster than Algorithm 508 and nearly as fast as the new implementation of the GibbsKingStockmeyer algorithm. (Author)
Descriptors : *Matrix theory, *Sparse matrix, Matrices(Mathematics), Algorithms, Reduction, Bandwidth, Wavefronts, Sizes(Dimensions), Errors, Profiles, FORTRAN, Coefficients, Symmetry, Heuristic methods
Subject Categories : Theoretical Mathematics
Distribution Statement : APPROVED FOR PUBLIC RELEASE