Installing Postgkyl

Installing Postgkyl is even easier than installing Gkyl. One option is to to use the Conda package manager, the other is to manually clone the Postgkyl repository. It should be noted that having both the repository and having Postgkyl installed through Conda on a single machine can cause unforeseen consequences.

Installing with Conda

The easiest way to install Postgkyl is to use the Conda installer that ships with Anaconda Python. Once you have installed Anaconda, the latest version of postgkyl can be installed from the gkyl channel using the command:

conda install -c gkyl postgkyl

Alternatively, one can permanently add the channel and install Postgkyl with:

conda config --add channels gkyl
conda install postgkyl

The big advantage of the installation through Conda is that all the dependencies are installed together with Postgkyl. To update the Conda installed version of Postgkyl, do:

conda update -c gkyl postgkyl

In order to use the Conda installation, the user requires to have the write permission to the Anaconda directory. This is not the case when anaconda is loaded as a module on a supercomputer. For supercomputers, one can either create a Conda environment or install the Anaconda into $HOME directory.

Install from Github

Unless you are planning develop postgkyl itself, you should install using Anaconda’s conda installer. Even if you plan to develop the code, it is a good idea to use Conda to install the dependencies:

  • adios
  • click
  • matplotlib
  • numpy
  • pytables
  • scipy
  • sympy

The postgkyl repository is private and access is only available to our close collaborators. This means, in practice, those who have funded projects or joint student projects with Princeton.

Once you have access to the Github repository, clone postgkyl into a new directory and add the path of the directory to your PYTHONPATH env variable. You will also need to add the pgkyl executable to your PATH.