I am trying to run a lithospheric extension model with free surface and passive particles to track the pT path of the upper crust but the simulation fails either during the initial mesh refinement at timestep 0 or at the start of timestep 1 when I remove the mesh refinement (first lines from the error file are below).
I am using version 2.3.0-pre
Thank you
Moh
An error occurred in line <3272> of file </nobackup/earmgo/aspect_workshop/candi/tmp/unpack/deal.II-v9.2.0/include/deal.II/dofs/dof_accessor.templates.h> in function
void dealii::DoFCellAccessor<DoFHandlerType, level_dof_access>::get_dof_values(const InputVector &, ForwardIterator, ForwardIterator) const [with InputVector = dealii::TrilinosWrappers::MPI::Vector, ForwardIterator = double *, DoFHandlerType = dealii::DoFHandler<3, 3>, bool level_dof_access = false]
The violated condition was:
this->is_artificial() == false
Additional information:
Can't ask for DoF indices on artificial cells.
Moh,
would you be willing to try this again with the current ASPECT development sources? Rene has done a lot of work in this area, and the problem may since have been fixed.
Best
Wolfgang
I’m also unsure exactly what the issue is here. The ability to use the particles with the free surface was reintroduced at some point in 2.3.0-pre, but that does not seem to be the issue.
I agree with Wolfgang’s suggestion to update to the most recent master branch and see if that resolves the issue.
In the next week or two I should have some tests done with extension models using active particles, and am happy to share those results when they have run successfully.
I tried to install the current ASPECT and ran into an issues:
[ 31%] Building CXX object CMakeFiles/aspect.dir/source/material_model/rheology/visco_plastic.cc.o
/home/home02/earmgo/safe/aspect_dir/aspect-2.4.0-pre/source/material_model/diffusion_dislocation.cc(452): error: namespace "std" has no member "make_unique"
diffusion_creep.parse_parameters(prm, std::make_unique<std::vector<unsigned int>>(n_fields));
^
/home/home02/earmgo/safe/aspect_dir/aspect-2.4.0-pre/source/material_model/diffusion_dislocation.cc(452): error: type name is not allowed
diffusion_creep.parse_parameters(prm, std::make_unique<std::vector<unsigned int>>(n_fields));
^
/home/home02/earmgo/safe/aspect_dir/aspect-2.4.0-pre/source/material_model/diffusion_dislocation.cc(456): error: namespace "std" has no member "make_unique"
dislocation_creep.parse_parameters(prm, std::make_unique<std::vector<unsigned int>>(n_fields));
^
/home/home02/earmgo/safe/aspect_dir/aspect-2.4.0-pre/source/material_model/diffusion_dislocation.cc(456): error: type name is not allowed
dislocation_creep.parse_parameters(prm, std::make_unique<std::vector<unsigned int>>(n_fields));
^
compilation aborted for /home/home02/earmgo/safe/aspect_dir/aspect-2.4.0-pre/source/material_model/diffusion_dislocation.cc (code 2)
make[2]: *** [CMakeFiles/aspect.dir/source/material_model/diffusion_dislocation.cc.o] Error 2
make[2]: *** Waiting for unfinished jobs....
make[1]: *** [CMakeFiles/aspect.dir/all] Error 2
make: *** [all] Error 2
It seems to be related to the compiler being c++11 instead of c++14, according to some posts that I found online, but I can’t figure out how to solve it! Any help would be appreciated .
we will be moving to requiring a c++ 14 compiler in about a week or so, so I don’t think we are going to go back to fix the issues you just reported. If you don’t want to upgrade, you can replace std::make_unique by std_cxx14::make_unique
There are maybe 6-7 places in source/ that needs fixing. I think that is all that is required to compile with c++11.
Thanks for the fix. It helped getting the compilation to 94% but now I am getting this error:
[ 94%] Building CXX object CMakeFiles/aspect.dir/contrib/world_builder/source/features/subducting_plate_models/temperature/linear.cc.o
/home/home02/earmgo/safe/aspect_dir/aspect-2.4.0-pre/unit_tests/parse_map_to_double_array.cc(87): error: name followed by "::" must be a class or namespace name
auto n_values_per_key = std_cxx14::make_unique<std::vector<unsigned int>>();
^
/home/home02/earmgo/safe/aspect_dir/aspect-2.4.0-pre/unit_tests/parse_map_to_double_array.cc(87): error: expected an identifier
auto n_values_per_key = std_cxx14::make_unique<std::vector<unsigned int>>();
^
/home/home02/earmgo/safe/aspect_dir/aspect-2.4.0-pre/unit_tests/parse_map_to_double_array.cc(87): error: expected a ";"
auto n_values_per_key = std_cxx14::make_unique<std::vector<unsigned int>>();
^
In any case, John shared with me one of his test setups using active particles that runs on the 2.3-branch.
I tried to install the current ASPECT and ran into an issues:
[ 31%] Building CXX object
CMakeFiles/aspect.dir/source/material_model/rheology/visco_plastic.cc.o
/home/home02/earmgo/safe/aspect_dir/aspect-2.4.0-pre/source/material_model/diffusion_dislocation.cc(452):
error: namespace “std” has no member “make_unique”
diffusion_creep.parse_parameters(prm, std::make_unique<std::vector<unsigned
I added the #include <aspect/compat.h> to the unit_tests/parse_map_to_double_array.cc but I still have the same issue (i.e., name followed by “::” must be a class or namespace name).
Concerning my initial issue with the particles, adding diffusivity to the mesh solved the issue. The model has reached 15 Myr so far on the 2.3-branch.
So, I am happy to wait for the cxx14 patch to be merged before compiling the master-branch.
Mohamed,
the patch I mentioned earlier has been merged. As for this particular file,
just remove everything from the file and leave it empty – you don’t need the
contents of the file.
.