Paolo Orsini <[email protected]> writes: > Thanks for this Jed, > > Any other alternatives you are aware of, for a parallel BILU(0) pre in > PETSc?
Not that I'm aware of, but there is no magic in blocks (it's purely implementation efficiency; if those entries are allocated in the matrix then scalar ILU is equivalent) so long as there aren't zero pivots.
signature.asc
Description: PGP signature
