Abstract
In this paper we examine ways of solving dense, banded systems on different parallel processors. We start with some considerations for processors with vector instructions, then discuss various algorithms for the solution of large, dense, banded systems on a parallel processor. We analyze the behavior of the parallel algorithms on distributed-storage architectures configured as rings, two-dimensional meshes with end-around connections (tori), boolean n-cube configured architectures, and bus-based and switch-based machines with shared storage. We also present measurements for two bus-based architectures with shared storage, namely, the Alliant FX/8 and the Sequent Balance 21000.
| Original language | English |
|---|---|
| Pages (from-to) | 219-246 |
| Number of pages | 28 |
| Journal | Parallel Computing |
| Volume | 5 |
| Issue number | 1-2 |
| DOIs | |
| State | Published - Jul 1987 |
| Externally published | Yes |
Funding
* Work supported in part by the Applied Mathematical Sciences subprogram of the Office of Energy Research, U.S. Department of Energy, under Contract W-31-109-Eng-38, and in part by the Office of Naval Research under Contract N00014-84-K-0043.
| Funders | Funder number |
|---|---|
| Office of Energy Research | |
| Office of Naval Research | N00014-84-K-0043 |
| U.S. Department of Energy | W-31-109-Eng-38 |
Keywords
- Alliant FX/8
- Sequent Balance 21000
- Solving large
- dense banded systems
- distributed-storage architectures
- parallel processors
- performance measurements