Abstract | ||
---|---|---|
The aim of this paper is the study of different methods for the solution of alternating-line problems, taking into account the evolution of architectural parameters on modern parallel computers, i.e. processors, memory hierarchy, and interconnection network performance. Three different kinds of solvers are studied: The Pipelined Gaussian Elimination scheme, the Matrix Transposition scheme, and the new one which is presented in this paper: the Mapping Transposition scheme, whose performance clearly betters, in many cases, that obtained by all the other methods, due to its better fitting to the characteristics of modern parallel computers. The experimental results have been obtained on a Cray T3E and on an SGI Origin 2000, up to 512 and 32 processors, respectively. |
Year | DOI | Venue |
---|---|---|
1999 | 10.1109/ICPP.1999.797406 | ICPP |
Keywords | Field | DocType |
alternating-line processes,matrix transposition scheme,alternating-line problem,different method,sgi origin,different kind,pipelined gaussian elimination scheme,interconnection network performance,cray t3e,mapping transposition scheme,modern parallel computers,modern parallel computer,parallel computer,difference equations,iterative methods,parallel processing,anisotropic magnetoresistance,gaussian elimination,partial differential equations,concurrent computing,mpi,adi,finite difference methods,differential equations | Memory hierarchy,Transpose,Computer science,Parallel processing,Parallel computing,Cray t3e,Gaussian elimination,Interconnection,Distributed computing,Network performance | Conference |
ISBN | Citations | PageRank |
0-7695-0350-0 | 4 | 0.59 |
References | Authors | |
10 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
D. Espadas | 1 | 23 | 2.69 |
Manuel Prieto | 2 | 349 | 31.23 |
Ignacio M. Llorente | 3 | 1174 | 65.01 |
Francisco Tirado | 4 | 404 | 32.26 |