[[Imported from SVN: r9509]]

[[Imported from SVN: r9508]]

This patch brings it back.

[[Imported from SVN: r9477]]

[[Imported from SVN: r9476]]

[[Imported from SVN: r9475]]

[[Imported from SVN: r9467]]

[[Imported from SVN: r9466]]

[[Imported from SVN: r9453]]

[[Imported from SVN: r9452]]

[[Imported from SVN: r9451]]

[[Imported from SVN: r9450]]

[[Imported from SVN: r9449]]

[[Imported from SVN: r9448]]

I can't make up my mind whether I want 'ctype' or 'field_type'

[[Imported from SVN: r9447]]

ADOL-C doesn't play nice with OpenMP.  Hence we have to disable
the latter for the time being.

[[Imported from SVN: r9446]]

[[Imported from SVN: r9445]]

When using the classes with ADOL-C, T becomes an 'adouble',
which is a non-trivial type.  However, the convexityField
is constexpr (and with a reason -- otherwise we could write
the initialization together with the declaration), and
constexpr and non-trivial types don't work together.

[[Imported from SVN: r9444]]

[[Imported from SVN: r9443]]

[[Imported from SVN: r9442]]

[[Imported from SVN: r9441]]

[[Imported from SVN: r9440]]

[[Imported from SVN: r9439]]

[[Imported from SVN: r9438]]

Because it doesn't work together with ADOL-C, claims the programmer
of ADOL-C.

[[Imported from SVN: r9437]]

[[Imported from SVN: r9436]]

And set the inner loop solver verbosity to QUIET

[[Imported from SVN: r9435]]

[[Imported from SVN: r9434]]

[[Imported from SVN: r9433]]

This number is nowadays infered from the number of transfer operators.

[[Imported from SVN: r9432]]

Since ADOL-C is not multi-threaded we will not be using OpenMP
in the foreseeable future.  And on a single-threaded code the
OpenMP timer and the Dune timer return the same values anyway.

[[Imported from SVN: r9431]]

[[Imported from SVN: r9430]]

Offer setInitialIterate instead

[[Imported from SVN: r9429]]

[[Imported from SVN: r9428]]

This still fails -- there must still be a bug in the Hessian computation

[[Imported from SVN: r9427]]

Because ADOL-C doesn't like calls to sqrt(0), obviously.
The use of sqrt(x) for small x can be avoided completely
using modified series expansions, because the overall
function exp is differentiable even at x=0.

[[Imported from SVN: r9426]]

Makes for a more precise Hessian matrix when using automatic differentiation.

[[Imported from SVN: r9425]]

[[Imported from SVN: r9424]]

[[Imported from SVN: r9423]]

[[Imported from SVN: r9422]]

[[Imported from SVN: r9421]]