Skip to content

Latest commit

 

History

History
63 lines (32 loc) · 3.2 KB

conda_setup.md

File metadata and controls

63 lines (32 loc) · 3.2 KB
Raw

How to setup a conda environment

Note: Hans has confirmed Dirac has a Miniconda install available. This allows users to install conda environments in their home directory which is accesible on compute nodes. For more info about the use of python and conda on the VSC, see this link.

Note: THIS PROCEDURE IS UNTESTED ON DIRAC. It has been used on VSC and in principle the workflow should be the same on Dirac. I have pointed out some VSC-specific steps.

(VSC-specific) Install Miniconda

  1. Get the executable for the latest Miniconda distribution :

    $ wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh

  2. Install Miniconda :

    $ bash Miniconda3-latest-Linux-x86_64.sh -b -p $VSC_DATA/miniconda3

    (generally advised to use $VSC_DATA instead of $VSC_HOME for installing conda environments due to storage quota - be careful since $VSC_SCRATCH is not permanent!)

  3. Add your Miniconda install your $PATH by adding the following line to .bashrc in $VSC_HOME :

    export PATH="${VSC_DATA}/miniconda3/bin:${PATH}"

    (It might be necessary to create .bashrc and obviously change ${VSC_DATA} based on the installation directory of Miniconda.)

Setup a Conda environment

Note: This is an example using gmx_MMPBSA. This workflow should also work for other conda installations, e.g. AmberTools.

Note: You may want to switch to the mamba solver. I have also found it to perform a lot better than the standard conda environment solver. Learn more in this link.

Note: at the time of writing the installation commands for dependencies of gmx_MMPBSA using conda & pip have already changed (especially version numbers of dependencies), see installation instructions for more info.

  1. Create the environment :

    $ conda create -n gmxMMPBSA python=3.9

    (This wil create a new environment with name gmxMMPBSA and use python3.9)

  2. Activate the environment :

    $ source activate gmxMMPBSA

  3. (VSC-specific) VSC is able to use multiple nodes for a single calculation. To optimize MPI parallelisation it is advised to use mpi4py from the intel repository :

    $ conda install -c intel mpi4py

  4. Install package dependencies :

    $ conda install -c conda-forge ambertools=21.12 compilers=1.2.0 -y -q

    $ python -m pip install git+https://github.com/Valdes-Tresanco-MS/[email protected]

    (The second command will install a modified version of parmed via pip in the conda environment since we have activated it previously.)

  5. (Optional - see gmx_MMPBSA installation instructions, link above) install gromacs :

    $ conda install -c conda-forge gromacs==2022.4 -y -q

  6. Install gmx_MMPBSA :

    $ python -m pip install git+https://github.com/Valdes-Tresanco-MS/gmx_MMPBSA

Hope this helps !! - CA