The ability to perform effective adaptive analysis has become a critical issue in the area of physical simulation. Of the multiple technologies required to realize a parallel adaptive analysis capability, automatic mesh generation is an enabling technology, filling a critical need in the appropriate discretization of a problem domain. The paving [1] algorithm's unique ability to generate a function-following quadrilateral grid is a substantial advantage in Sandia's pursuit of a modified h-method adaptive capability. This characteristic combined with a strong transitioning ability allow the paving algorithm to place elements where an error function indicates more mesh resolution is needed. Other desirable characteristics of this algorithm include its boundary sensitivity and orientation insensitivity (elements near the boundary are of the highest quality and the spatial orientation of the geometry has no effect on the resulting mesh). Although the original paving algorithm is highly ser...

In this paper we define a knowledge discovery in databases (KDD) methodology to automatically generate metadata (i.e., knowledge rules) from software/machine pair performance databases. This metadata can be used to characterize the computational behavior of various classes of software or machines. The core and the most computationally intensive part of the KDD methodology is the data mining phase which identifies "interesting" patterns from the performance data. The discovery patterns are expressed in a high level representation to be used to summarize and predict the computational behavior of the targeted software/machine. This paper presents an implementation and evaluation of the proposed KDD process for a class of scientific software together with three data mining algorithms (ID3, HOODG, and CN2). For this case study we have selected a set of software that implements the "mesh/grid partitioning " phase of the domain decomposition approach used for the parallel processing of partial differential equation (PDEs) computations. The raw performance database is generated from a population of elliptic PDEs and PELLPACK [HRW 98] solvers by varying the PDE domain, mesh, and domain partitioning (DP) algorithm. The goal of the KDD process here is to evaluate the performance of PELLPACK, CHACO [HL95c], METIS [KK95c], and PARTY [PD96] algorithms/software. This case study shows that (a) the three data mining algorithms used are qualitatively and quantitatively equally effective, (b) the knowledge discovered for the DP algorithms by this KDD process is quantitatively similar to that deduced by purely experimental observations [VH97], and (c) the KDD process is not limited by the size of the performance data and its dimensionality.

In this chapter, we present a new approach for parallel generation and partitioning of 3-dimensional unstructured meshes. The new approach couples the mesh generation and partitioning problems into a single optimization problem. Traditionally these two problems are solved separately with I/O and data movement overheads that exceed 90% of the total execution time for generating, partitioning, and placing very large meshes on distributed memory parallel computers. The new approach minimizes the I/O and data-movement overheads by eliminating redundant memory operations (loads/stores) from and to cache, local & remote memory, and discs. Our preliminary results show that the new approach is nine times faster than the traditional approach for generating, partitioning, and distributing very large 3-dimensional unstructured meshes.

Paving [1] is an automated mesh generation algorithm which produces all-quadrilateral elements. It can additionally generate these elements in varying sizes such that the resulting mesh adapts to a function distribution, such as an error function. While powerful, conventional paving is a very serial algorithm in its operation. Parallel paving is the extension of serial paving into parallel environments to perform the same meshing functions as conventional paving only on distributed, discretized models. This extension allows large, adaptive, parallel finite element simulations to take advantage of paving's meshing capabilities for h-remap remeshing. A significantly modified version of the CUBIT [2] mesh generation code has been developed to host the parallel paving algorithm and demonstrate its capabilities on both two dimensional and three dimensional surface geometries and compare the resulting parallel produced meshes to conventionally paved meshes for mesh quality and algorithm perf...