pymistral

Effective use of python with xarray and dask to analyse Earth-System-Modelling output on the mistral supercomputer at DKRZ.

Examples

Check out our examples:

• easy access via intake-esm on CMIP5, CMIP6 and MiKlip output on mistral
• easy access via intake to ICDC observations on mistral
• grid handling of MPIOM via xgcm
• plotting the curvilinear MPIOM with cartopy: xr.DataArray.plot_map()

Contact

• Sebastian Milinski (sebastian.milinski@mpimet.mpg.de)

• Aaron Spring (aaron.spring@mpimet.mpg.de)

System configuration

The following steps are necessary to use pymistral.

Environment

Create a new conda environment using the pymistral.yml file. This has to be done on the supercomputer when using pymistral on an HPC system or locally for a standalone configuration.

conda env create -f ci/pymistral.yml

If the environment already exists and needs to be updated, use:

source activate pymistral
conda env update -f=ci/pymistral.yml

Script to start jupyterlab on HPC

This part is only necessary when jupyterlab and pymistral is used on a remote HPC system. A detailed explanation is at the DKRZ website: https://www.dkrz.de/up/systems/mistral/programming/jupyter-notebook

pymistral_preload

Place the jupyter_preload file into your home directory on mistral and change the conda environment name if necessary. Alternatively, you can create your own conda: https://docs.conda.io/projects/conda/en/latest/user-guide/install/linux.html Here don't use your $HOME on mistral, specify path like /work/yourgroup/m??????/miniconda3.

./start-pymistral

Adjust the start-jupyter script with your username and project number.