A Hierarchical Partition Model for Parallel Adaptive Finite
Element Computation
Defense: Tuesday, July 25, 2000, 10:00 AM, Lally 104
Examining Committee:
Joseph E. Flaherty, Thesis Adviser
Mark S. Shephard, Member
Boleslaw K. Szymanski, Member
Franklin T. Luk, Member
Abstract:
Software tools for the solution of partial differential equations
using parallel adaptive finite element methods have been developed.
We describe the design and implementation of the parallel mesh
structures within an adaptive framework. In a parallel adaptive
computation, the mesh changes during the computation, necessitating a
dynamic redistribution of data. Mesh structures must not only support
the adaptivity, but this dynamic redistribution. The distributed mesh
structures are built upon a serial mesh database, and our enhancements
to that system are described. This work began with several
enhancements to a parallel mesh database developed previously. Our
redesigned distributed mesh structures are based on the
object-oriented and generic programming paradigms. The most
fundamental concept is that of a hierarchical partition model used to
distribute finite element meshes and associated data on a parallel
computer. The hierarchical model represents heterogeneous processor
and network speeds, and may be used to represent processes in any
parallel computing environment, including an SMP, a distributed-memory
computer, a network of workstations, or some combination of these. We
describe several example applications, examining both the solution
properties and the effectiveness of several algorithms for dynamic
load balancing. We present experiments which show that the
information about different processor speeds, memory sizes, and the
corresponding interconnection network can be useful in a dynamic load
balancing algorithm which seeks to achieve a good balance with minimal
interprocessor communication penalties when a slow interconnection
network is involved. Examples of how such architecture-dependent load
balancing can be applied are also given.
Dissertation:
The full text of the dissertation is available below.
