Changes to default Python environment on ARCHER

As part of our continuing effort to improve and clarify the ARCHER Service we are planning to update the way in which Python works by default for users on Wednesday 30 September 2016.

Currently, all users have the Anaconda Python environment loaded by default when they log in. This can cause confusion and unexpected behaviour for users using Python who are not expecting to access Anaconda.

To simplify the ARCHER Python setup we will be changing the service so that none of the Python modules (anaconda or python-compute) are loaded by default. Users who wish to access these modules will have to load them on the command line or in their job submission scripts.

Python will still be available in user environments at login via the default system-installed Python (2.6.9).

Choosing a Python environment on ARCHER

The first step in using Python on ARCHER is deciding which is the most appropriate set of modules to use. The easiest way to decide this is based on which nodes you will be running on:

• If you are going to be running Python on the login nodes or the PP nodes then you should use the Anaconda distribution (module load anaconda).
• If you are running Python on the compute nodes (via aprun) then you should use the native distribution (module load python-compute).

More information on using Python on ARCHER can be found in the Using Python on ARCHER chapter of the ARCHER User Guide.

If you have any questions on this change, please contact the ARCHER Helpdesk.