PARALLEL FREQUENCY FILTERING

The frequency filtering method is a robust and efficient ILU-like solv er for large sparse systems (cf. [9,10]). Combining this method with t he so-called Schur-complement DD method, we obtain a fast parallel sol ver. In this context, frequency filtering can be applied as solver ins ide the subdomains as well as for the treatment of the arising Schur c omplements. Especially for those, the method is well suited since it i s highly parallelizable by recursively applying the same decomposition as to the original system. In this paper, an implementation of the fr equency filtering domain decomposition (FFDD) method on a multiprocess or system will be presented and the numerical results of some variants thereof be discussed. The scaling behaviour of the algorithm for an i ncreasing number of processors is almost optimal.