Wednesday, May 12

Recent Advances in Sparse Matrix Techniques

(Invited Minisymposium)

4:00 PM-6:00 PM
Room: Atlanta 2

This minisymposium concerns the development of scalable sparse solvers, including direct methods (Gaussian elimination and its variants) and iterative methods (conjugate gradients and domain decomposition). The speakers will discuss a hybrid solver using incomplete Cholesky preconditioners; a grid-based object oriented framework for developing scalable solvers; and scalable alternatives to traditional approaches for partial pivoting and triangular solves.

Organizer: Padma Raghavan
University of Tennessee, Knoxville

4:00-4:25 Structural Results for Sparse Orthogonal Factorization
Esmond G. Ng, Oak Ridge National Laboratory
4:30-4:55 An Object-Oriented Framework for Grid Solvers
Akhil Vidwans, University of Illinois, Urbana-Champaign
5:00-5:25 Making Sparse Gaussian Elimination Scalable by Static Pivoting
Xiaoye S. Li, NERSC, Lawrence Berkeley Laboratory
5:30-5:55 Scalable Schemes for Repeated Sparse Triangular Solves
Padma Raghavan, Organizer

AN99 Home


Program Updates

Speaker Index




LMH, 1/18/99, MMD, 1/26/98