- recent GCC version produce error when "-isystem /usr/include" is added
  to command line. Error is: '#include_next <math.h>' not found
- the removed "hack" added the cuda include dir with "-isystem" to
  the CUDA compiler arguments to prevent warnings from boost header
  -> currently there seem to be no warnings generated from boost headers
     so "hack" is removed

added filter to exclude external libraries

- NearestNeighborFieldInterpolator::get needs to break out of all loops, just not the current one. While the test seemed to succeed anyway, it crashed with corrupted memory during exit when compiled with Intel 17 in Release mode.
- FftTest should not use thread-unsafe RNG in parallel section. Otherwise the Intel 16 compiler crashes.
- Revert to -openmp on Intel < 16.0.3, which crashes while processing some of our OpenMP loops (e.g. BoundaryHandling::treatDirection).

Also adds OpenMP support for macOS (requires Xcode 7.0, CMake 3.12 and separate libomp from MacPorts/Homebrew)

- fix find_package(MPI) when WARNING_ERROR is defined and recent version of CMake is used
- fix python_coupling Manager::addPath

Michael Kuron authored 7 years ago and Christian Godenschwager committed 7 years ago

- Need to find Boost library path before we search for the Python libraries, even if we don't need any other libraries
- std::experimental::filesystem::current_path is broken in libstdc++ 5.4 and lower

Michael Kuron authored 7 years ago and Christian Godenschwager committed 7 years ago

API changes in blockforest::PhantomBlockForest, blockforest::loadbalancing, pe::amr::weight_assignment
mpi can now only send/receive walberla::optional instead of boost::optional

Michael Kuron authored 7 years ago and Christian Godenschwager committed 7 years ago

API change for vtk::VTKOutput::getFilenames

OpenMPI 3.0 by default does not allow running with more processes than CPUs

Also added wrapper for Metis & ParMetis to prevent it from polluting the global name space

- normal CMake Python search can be used if the python interpreter is
  searched before the searching the libraries

- nvcc also printed warnings in system libs like boost

FindCUDA apparently can't handle CMake "generator expressions"
these are removed before calling FindCUDA and restored afterwards

- default python is the one reported by python3-config

two versions available now via CMake switch:
   - by default the Python module is built without LBM module, since
     lbmpy can be used for this. Reduces compile time substantially.
   - if the LBM is needed in python, for example for the lbmpy testsuite
     it can be included by setting the CMake switch
     WALBERLA_BUILD_WITH_PYTHON_LBM