Performance and scalability of finite element analysis for distributed parallel computation (Q1328092)
From MaRDI portal
scientific article
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | Performance and scalability of finite element analysis for distributed parallel computation |
scientific article |
Statements
Performance and scalability of finite element analysis for distributed parallel computation (English)
0 references
2 January 1995
0 references
Parallelization of a \((h,p)\) finite element code by domain decomposition and iterative solution of the linear equations by the conjugate gradient method. The iteration is based on the element matrices, no global stiffness matrix is assembled. The increased storage overhead is mentioned, but not the computational overhead. For a 2D partial differential equation a model for the time behavior is established and tested for the nCUBE2 up to 512 processors. With this model scaling for different objective functions is investigated.
0 references
parallel computation
0 references
finite element method
0 references
\(h,p\) method
0 references
performance
0 references
domain decomposition
0 references
conjugate gradient method
0 references
