Abstract
This paper describes algorithms for solving narrow banded systems and the Helmholtz difference equations that are suitable for multiprocessing systems. The organization of the algorithms highlight the large grain parallelism inherent in the problems.
| Original language | English |
|---|---|
| Pages (from-to) | 223-235 |
| Number of pages | 13 |
| Journal | Parallel Computing |
| Volume | 1 |
| Issue number | 3-4 |
| DOIs | |
| State | Published - Dec 1984 |
| Externally published | Yes |
Funding
Let the linear system under consideration be denoted by Ax =f (1) where A is a banded diagonally dominant matrix of order n. We assume that the number of superdiagonals m << n is equal to the number of subdiagonals. On a sequential machine such a system would be solved via Gaussian elimination without pivoting at a cost of O(m2n) arithmetic operations. We describe here an algorithm for solving this system on a multiprocessor of p processing units. Each unit may be a sequential machine, a vector machine, or an * Work supported in part by the Applied Mathematical Sciences Research Program (KC-04-02) of the Office of Energy Research of the U.S. Department of Energy under contract W-31-109-Eng-38. ** Work supported in part by the National Science Foundation under grant US NSF MCS 81-17010.
| Funders | Funder number |
|---|---|
| Office of Energy Research | |
| National Science Foundation | US NSF MCS 81-17010 |
| U.S. Department of Energy | W-31-109-Eng-38 |