No label defined (Q4035287)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 176183
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 176183 |
Statements
18 May 1993
0 references
quadrant interlocking factorization
0 references
SIMD-parallel computers
0 references
adaptive method
0 references
banded linear systems
0 references
finite element
0 references
extendable sparse factorization
0 references
parallel linear solver
0 references
0.8472150564193726
0 references
0.8383033871650696
0 references
0.8211595416069031
0 references
0.803667426109314
0 references
[For the entire collection see Zbl 0753.00013.] The quadrant interlocking factorization (QIF) method, suggested by the first author and \textit{M. Hatzopoulos} [ Int. J. Comput. Math. 7, 227-238 (1979; Zbl 0442.65019)] for the solution of nonsingular linear systems \(Au=s\) on SIMD-parallel computers is considered. An adaptive method is suggested for the coefficient matrix \(A\) being a nonsingular banded diagonally-dominated matrix. The authors apply the QIF method to the numerical solution of banded linear systems, which occur in the finite element discretization of partial differential equations. Based on the extendable sparse factorization techniques of the authors [Comput. 24, 325-339 (1980; Zbl 0428.65056)], the parallel linear solver can be used either as a direct or an iterative method, where the sparse factorization is used in conjunction with implicit schemes. According to the theoretic evaluation, the method requires the number of \(\max\{2(r_ 1+r_ 2)^ 2,\;2(r_ 1+r_ 2+1)\}\) parallel processors where \(r_ 1\) and \(r_ 2\) are the outer most off-diagonal terms.
0 references