Parallel elliptic solvers and their application to oil recovery simulation

摘要

The paper is devoted to investigations of algorithms for solving 2D elliptic PDEs of the second order on parallel computers with distributed memory. The main attention is paid to the original iterative (α - β)-algorithm. Its parallel variants are analysed and empirically compared with Gauss-Seidel, SOR, ad hoc SOR and the generalized alternating triangular conjugate gradient method. Test predictions as well as results of solving the applied oil recovery problem by these methods are presented: convergence rates, CPU times and parallelization efficiencies are estimated. This work is supported by Russian Foundation for Basic Research (grant No. 99-01-01215).