Chris Sewell

> What's the best way to get around this? Off-hand, I guess the easiest solution is just to delete the generated `roles/marvel-nccr.aiida` folder > I haven't tried it locally so...

If possible @borellim, It would be good if you could try narrowing down what ones need to be added, e.g. a classic bisection approach; add half of them to tox.ini,...

Another interesting repo that contains a slurm role is: https://github.com/NVIDIA/deepops

See https://github.com/marvel-nccr/ansible-role-conda-codes/pull/1; currently some issues to address on how to activate the Conda environments programmatically

For reference these are the longest running tasks in the current (v16.04.21a) vagrant build ``` =============================================================================== marvel-nccr.fleur : Make fleur executables ---------------------------- 940.93s marvel-nccr.bigdft : Compile bigdft executables ----------------------- 688.38s...

I guess another thing to bare in mind is "access" to the executables for users. Currently, users can simply run e.g. `mpirun pw.x`, whereas now you would first need to...

> Since the activation of the environment is not necessary for the executables, we could play around with symlinks Do the executables need to use the `mpirun` executable from their...

> In the long term, ideally we should aim for maximum compatibility in the conda packages as well - I guess at some point there will always be packages that...

> i.e. if abinit was using mpich also in the QM It doesn't, everything is compiled against the same openmpi (from apt)

Well abinit does not work quite correctly at present, that's the problem. See https://github.com/marvel-nccr/ansible-role-abinit/issues/14