Vectorization of algorithms for solving systems of difference equations
Today's fastest computers achieve their highest level of performance when processing vectors. Consequently, considerable effort has been spent in the past decade developing algorithms that can be expressed as operations on vectors. In this paper two types of vector architecture are defined. A discussion is presented on the variation of performance that can occur on a vector processor as a function of algorithm and implementation, the consequences of this variation, and the performance of some basic operators on the two classes of vector architecture. Also discussed is the performance of higher-level operators, including some that should be used with caution. With both types of operators, the implementation of techniques for solving systems of difference equations is discussed. Included are fast Poisson solvers and point, line, and conjugate-gradient techniques. 1 figure.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 6893475
- Report Number(s):
- LA-UR-81-270; CONF-810113-1
- Resource Relation:
- Conference: Simulation methods for nuclear power systems conference, Tucson, AZ, USA, Jan 1981
- Country of Publication:
- United States
- Language:
- English
Similar Records
Vectorization of algorithms for solving systems of difference equations
Implementing techniques for elliptic problems on vector processors