[Getdp] PETSC --with-64-bit-indices=1

[Mark van Doesburg]

	Hi Mark - Thanks: I merged the patch.

Thanks.

	> I'll try to port the Parallelization patch by Davide Rondini to the
	> latest snapshot.

I'm going to reply to myself on this one. I found out the mumps solver
seems to be the most efficient on the problems I've encountered so
far. This algorithm doesn't need the parallelization patch, so I'm
putting that on hold for now.

Mark van Doesburg.