Parallel pivoting algorithms for sparse symmetric matrices (Q760156)

It is shown that for sufficiently sparse symmetric matrices some pivots may be processed simultaneously. The author extends the results of \textit{J. A. G. Jess} and \textit{H. G. M. Kees} [IEEE Trans. Comput. C-31, 231-239 (1982; Zbl 0479.68034)] by removing the no fill-in restriction. Algorithms which yield maximum parallelism, are given for a computer with sufficiently many processors and a common memory. Additional algorithms are given for a computer with an arbitrary but fixed number of processors. It is assumed that the numerical stability does not depend on the order in which the diagonal elements are chosen as pivots.