I've just readed Matt mail about an implementation for using schur complement system.
The actual code is inside petsc4py svn repo at: http://petsc4py.googlecode.com/svn/trunk/petsc/lib/ext/petsc/src/ksp/pc/impls/ There are a dir named 'schur', that implementation required a global MATIS. There are another dir named 'schur_aij' wich is an attemp of implementing the same for global MATMPIAIJ matrices, however this has not been tested very much. All this is being using here at CIMEC. I've just posted in petsc-users a basic description of the (rather involved) implementation. Take a look in case any of you is insterested. I've never decided to propose it for inclusion in PETSc because never felt confident all this really make sense and did not want to contaminate PETSc with a PC implementation that I do not know will be useful outside our particular use case. Regards, -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594
