A Dynamically Partitioned Out-of-Core Skyline Solver for Micro-Computers

A J du Toit and W S Doyle
Additional contact information
A J du Toit: University of Cape Town, Department of Civil Engineering
W S Doyle: University of Cape Town, Department of Civil Engineering

Chapter S14 in Numerical Techniques for Engineering Analysis and Design, 1987, pp 649-658 from Springer

Abstract: Summary In this paper we present improvements to the well known block partitioned form for the out-of-core solution of sets of linear equations. Our solver operates on a coefficient matrix which is stored out-of-core as a one dimensional array of values under the skyline. The improvements pertain to the partitioning of the coefficient matrix into blocks of reduced and unreduced terms which vary as the decomposition of the matrix proceeds. This dynamic form of partitioning selects an optimum size of the blocks and leads to substantial savings in terms of execution time and data transfer from backup to storage. The algorithm uses the Cholesky method for decomposition of the coefficient matrix but can be modified to accommodate Crout reduction for instance. The solver was developed specifically for the micro-computer environment but the concepts are general and may be implemented on a full range of machines.

Keywords: Coefficient Matrix; Load Vector; Block Partitioning; Fixed Block; Dynamic Partitioning (search for similar items in EconPapers)
Date: 1987
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:sprchp:978-94-009-3653-9_72

Ordering information: This item can be ordered from
http://www.springer.com/9789400936539

DOI: 10.1007/978-94-009-3653-9_72

Access Statistics for this chapter

More chapters in Springer Books from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().