Treffer: A fully portable high performance minimal storage hybrid format Cholesky algorithm

Title:
A fully portable high performance minimal storage hybrid format Cholesky algorithm
Authors:
Fred G. Gustavson, John Reid, Jerzy Waśniewski, Bjarne Stig Andersen, John A. Gunnels
Source:
ACM Transactions on Mathematical Software. 31:201-227
Publisher Information:
Association for Computing Machinery (ACM), 2005.
Publication Year:
2005
Subject Terms:
Complexity and performance of numerical algorithms, LAPACK, 0202 electrical engineering, electronic engineering, information engineering, Packaged methods for numerical algorithms, 02 engineering and technology, 0101 mathematics, Direct numerical methods for linear systems and matrix inversion, LINPACK, symmetric positive-definite dense systems, 01 natural sciences, Cholesky factorization
Document Type:
Fachzeitschrift Article
File Description:
application/xml
Language:
English
ISSN:
1557-7295
0098-3500
DOI:
10.1145/1067967.1067969
Access URL:
http://orbit.dtu.dk/ws/files/5875386/imm3173.pdf
https://orbit.dtu.dk/en/publications/a-fully-portable-high-performance-minimal-storage-hybrid-format-cholesky-algorithm(73b4f63e-6036-4fae-b44a-3bbadb1eba41).html
https://dblp.uni-trier.de/db/journals/toms/toms31.html#AndersenGGRW05
https://core.ac.uk/display/13755213
http://www.scd.stfc.ac.uk/SCD/resources/reports/aggrwRAL2004017.pdf
https://dl.acm.org/doi/10.1145/1067967.1067969
Rights:
URL: https://www.acm.org/publications/policies/copyright_policy#Background
Accession Number:
edsair.doi.dedup.....0e05a1d879b18e8be4a7cade878ffb6f
Database:
OpenAIRE

Weitere Informationen

We consider the efficient implementation of the Cholesky solution of symmetric positive-definite dense linear systems of equations using packed storage. We take the same starting point as that of LINPACK and LAPACK, with the upper (or lower) triangular part of the matrix stored by columns. Following LINPACK and LAPACK, we overwrite the given matrix by its Cholesky factor. We consider the use of a hybrid format in which blocks of the matrices are held contiguously and compare this to the present LAPACK code. Code based on this format has the storage advantages of the present code but substantially outperforms it. Furthermore, it compares favorably to using conventional full format (LAPACK) and using the recursive format of Andersen et al. [2001].