The first annual large dense linear system survey
- 1 October 1991
- journal article
- Published by Association for Computing Machinery (ACM) in ACM SIGNUM Newsletter
- Vol. 26 (4) , 6-12
- https://doi.org/10.1145/122645.122648
Abstract
In the March 24, 1991 issue of NA Digest, I submitted a questionnaire asking who was solving large dense linear systems of equations. Based on the responses, nearly all large dense linear systems today arise from either the benchmarking of supercomputers or applications involving the influence of a two dimensional boundary on three dimensional space. Not surprisingly, the area of computational aerodynamics or aero-electromechanics represents an important commercial application requiring the solution of such systems. The largest unstructured matrix that has been factored using Gaussian Elimination was a complex matrix of size 55,296. The largest dense matrix solved on a Sun using an iterative method was a real matrix of size 20,000. It is unclear at this time whether dense methods are truly needed at all for huge matrices. It is intended to survey users every year with the hope of including more applications as I am made aware of them.Keywords
This publication has 6 references indexed in Scilit:
- Sparse Approximation for Solving Integral Equations with Oscillatory KernelsSIAM Journal on Scientific and Statistical Computing, 1992
- Sparse matrix approximation to an integral equation of scatteringCommunications in Applied Numerical Methods, 1990
- Origin and development of the method of moments for field computationIEEE Antennas and Propagation Magazine, 1990
- Impedance matrix localization produces a sparse moment method matrixPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1990
- Panel Methods in Computational Fluid DynamicsAnnual Review of Fluid Mechanics, 1990
- Basic Principles and ApplicationsPublished by Springer Nature ,1984