Deganit Armon

Learn More
Nested dissection (ND) is a method for solving sparse n n linear systems eeciently by imposing an order of elimination on the variables. Parallel nested dissection (PND) algorithms, which do this variable elimination in parallel, have been developed for the PRAM model and for grid architectures. In this paper, we describe the following improvements to the(More)
Our work is based on the pioneering work in sphere separators done by Miller, Teng, Vavasis et al, [8, 12], who gave efficient static (fixed input) algorithms for finding sphere separators of size s(n) = O(n d−1 d) for a set of points in R d. We present dynamic algorithms which maintain separators for a dynamically changing graph. Our algorithms answer(More)
This paper is concerned with algorithms for the efficient parallel solution of sparse, symmetric n × n linear systems via direct factorization, which eliminate groups of variables in stages. The algorithms make use of nested dissection to recursively cut data into pieces, producing orderings for direct elimination of the variables, so to reduce the storage(More)
  • 1