Skip to content

Latest commit

 

History

History
65 lines (49 loc) · 2.1 KB

COMPILE.md

File metadata and controls

65 lines (49 loc) · 2.1 KB

Underworld 2 Compilation

Note that for most personal computer usage, we recommend using Underworld through a Docker container environment. See README.md for details.

For HPC usage, you will generally need to compile Underworld, though we support Shifter where available (check with the HPC admins). Basic compilation information is below, though please submit a github issue if you have have difficulties.

Dependencies

  • MPI
  • PETSc
  • numpy
  • swig
  • h5py

Getting the code

git clone https://github.com/underworldcode/underworld2.git

Compilation

We periodically post build recipes for particular platforms on the underworld blog but the build strategy generally looks like this

    $ cd libUnderworld
    $ ./configure.py
    $ ./compile.py

Check available configuration options using ./configure.py --help.
Library testing options may be found by running ./scons.py --help.

Self compiled Underworld2 usage

You will first need to make the project directory available to import within python:

    $ export PYTHONPATH=$PYTHONPATH:/top/directory/of/project

(note that if you are not using the bash shell, the required command will be different.)

h5py-mpi

Underworld now supports non-mpi versions of h5py. Note however that for large parallel simulations, saving data to disk may become a bottleneck, and collective IO via MPI-enabled h5py is recommended. To install h5py with MPI enabled, you will first required a parallel enabled version of libhdf5. The following command may be useful for installed MPI-enabled h5py:

    $ CC=mpicc HDF5_MPI="ON" HDF5_DIR=/path/to/your/hdf5/install/ pip install --no-binary=h5py h5py

or alternatively you might use CC=h5pcc (if available). Please check the h5py site for more information. Underworld will automatically perform save()/load() operations collectively if MPI-enabled hdf5 is available.