[name removed] users and developers,
I spent some times looking how to package [name removed] for conda and it’s now done !
Package description:
2 configs are built : gnu.static and gnu.static.debug taken from https://gitlab.pam-retd.fr/otm/wintel/-/blob/main/config/systel.cfg.
Support parallel computation.
Require 1 extra package which is not in default repo : msmpi-compilers (providing mpif90)
Performance should be similar to wintel install
I consider this package as an alpha for now.
Only t2d_malpasset-fine.cas in parallel has been tested.
Features
- Parallel computation
- Switch debug mode
telemac-debug on/off
Not implemented (yet…)
- Dynamic build (I have trouble with API compilation
- HDF5 and MED support
- MUMPS (openblas, scalapack) support
- aed2 support
- gotm support
- Provide examples in separate package
- Provide documentation in separate package
How to install?
- Install miniforge3: https://github.com/conda-forge/miniforge#miniforge3
- In miniforge prompt, run :
mamba create -n telemac-mascaret -c nicogodet telemac-mascaret
mamba activate telemac-mascaret
python -m config
python -m telemac2d cas_file.cas
[name removed], Scipy and Matplotlib are installed (marked as dependecies). Other python packages (listed in https://gitlab.pam-retd.fr/otm/wintel/-/blob/main/python_requirements.txt) should be manually installed (not tested)
Package recipe
Github repo: https://github.com/nicogodet/telemac-mascaret-feedstock
Anaconda: https://anaconda.org/nicogodet/telemac-mascaret
Via this topic, I call for testers, developers and maintainers to provide a complete and efficient telemac package