Pipelined iterative methods for shared memory machines (Q1825591)
From MaRDI portal
scientific article
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | Pipelined iterative methods for shared memory machines |
scientific article |
Statements
Pipelined iterative methods for shared memory machines (English)
0 references
1989
0 references
This paper describes a parallel pipelining technique applied to the SOR method for a banded matrix. The basic idea is that several processors can produce several new iterates, with the indices of the variables they work on separated at least by the bandwidth. Computational experience with an implementation on a 28 processor shared memory sequent symmetry system is reported.
0 references
parallel computation
0 references
parallel pipelining
0 references
SOR method
0 references
banded matrix
0 references
shared memory sequent symmetry system
0 references
