the ThreadManager from ebos was not called which resulted in some
havoc when attempting multi-threaded runs.
v2: use opm_get_max_threads() directly. thanks to [at]akva2 for the heads-up.
it seems like most build systems pass a -DHAVE_CONFIG_H flag to the
compiler which still causes `#if HAVE_CONFIG_H` to be false while it
clearly is supposed to be triggered.
That said, I do not really see a good reason why the inclusion of the
`config.h` file should be guarded in the first place: the file is
guaranteed to always available by proper build systems, and if it was
not included the build either breaks at the linking stage or -- at the
very least -- the runtime behavior of the resulting libraries will be
very awkward.
3x3 matrix block inversion in FieldMatrix is numerically unstable
including version 2.5.0. Therefore the previous if clause was wrong
as it activated the use of FieldMatrix already for 2.5.0 (the version
in Debian stable). With this commit we use MatrixBlock for version 2.5.0.
since I has not figure out a way to detect/catch the singularity of the
matrix with UMFPack, currently, we only check the validity of the
results to make sure inf or nan not enter other part of the simulation.
It can easiliy results in assertion failure, and causes the running to
be terminated. After all, it mostly likely is a numerical issue
generated during the non-linear iteration process.
When there is not ms well involved, all the ms well related is not
initialized. It causes problem when we want to add ms well after some
time of running.
The bug fix the running with model 2.
for standards wells, all the three equations are mass balance
equatiions, it is safe to use numWellEq.
for MS wells, there is one extra pressure equation, it should be the
number of mass balance equations. If we do not put the polymer equation
inside the well equations, we will face dilemma soon.
to determine whether we will treat mutlisegment wells with
MultisegmentWell well model. Currently, it will be faulse by default.
It might be changed to be true by default when considering the
multisegment well model is well tested.
It can be something rather tricky. It looks like with simple trival
initization, it is difficutl to converge. But when there is change to
the well, direct copying will not work anymore. Will we also facing
convergence problem?
