New factorization codes for sparse,symmetric and positive definite matrices

Suitable techniques for storing the matrix pattern during the factorization of sparse, symmetric and positive definite matrices are considered. Especially we discuss the consequences of switching from a sparse factorization code to a full code when the uneliminated part of the matrix is full or almost full. The resulting codes seem to be among the most efficient for solving one-off problems regarding both execution time and storage requirements.This work has been supported by the Danish Natural Science Research Council, Grant No. 511-8476.