Performance and Scalability of Finite Element Analysis
for Parallel Computation
- E. Barragy
- Department of Engineering Mechanics
- University of Texas
- Austin, TX 78712
- G. F. Carey
- Department of Engineering Mechanics
- University of Texas
- Austin, TX 78712
- R. A. van de Geijn
- Department of Computer Sciences
- University of Texas
- Austin, TX 78712
Abstract
A two-dimensional (h,p) finite element scheme
for distributed parallel computation is developed. The approach
is based on an element-by-element domain decomposition and is
implemented on the nCUBE2 system. Example problems are used to
demonstrate performance of the algorithm for a range of (h,p)
and to validate the run-time model. The run-time model is then
used to examine the scaling properties of conjugate gradients and
the (h,p) FEM over a range of p.
E. Barragy, G. F. Carey, and R. van de Geijn,
``Performance and Scalability of Finite Element Analysis
for Parallel Computation,''
Journal of Parallel and Distributed Computing , 21, pp. 202-212 (1994).