Skip to content

Latest commit

 

History

History
86 lines (63 loc) · 3.19 KB

02-setup_environment.md

File metadata and controls

86 lines (63 loc) · 3.19 KB

Next: Prepare and run simulation, Previous: Accessing the Cirrus cluster, Up: Main page

Setting up the environment

Obtaining all resources

Before you start, you must clone this repository to your working directory on Cirrus. You can do this by using the command:

cd /work/tc066/tc066/$USER
git clone https://github.com/EPIC-model/workshop.git

How to load the EPIC environment

In order to simplify your task, we have pre-installed a working EPIC executable. After logging into Cirrus, you can load the environment with the following commands:

module use /work/tc066/tc066/shared/modules
module load epic

The epic module will automatically load all the other required modules (gcc, mpi, hdf5, and netcdf). You can confirm which modules you have loaded with the module list command:

tc066-rfga@cirrus-login2:~$ module list
Currently Loaded Modulefiles:
 1) git/2.37.3   3) /mnt/lustre/e1000/home/y07/shared/cirrus-modulefiles/epcc/setup-env   5) gcc/10.2.0(default)      7) hdf5-epic     9) epic
 2) epcc/utils   4) htop/3.2.1                                                            6) openmpi/4.1.6(default)   8) netcdf-epic

If you would like to test the installation of EPIC, you can try out this build script. For this purpose, you would not load the EPIC module (last command). If you want to try on your own laptop, you may have noticed that EPIC requires a working gfortran compiler, MPI and a netCDF (requires HDF5) installation. To inspect the netCDF output files while a simulation is running, you can additionally load the ncview module with

module load ncview-epic

To open an image file, load the ImageMagick module, and use the display command.

module load ImageMagick

Note

Your ssh connection must have been made with the flag -X in order to run ncview or ImageMagick's display.


How to load the Python virtual environment

You can prepare the input to EPIC and analyse its output using our tools written in Python. For this purpose, you first need to load the virtual environment by typing the subsequent commands:

module load python/3.9.13
CONDA_ROOT=/work/tc066/tc066/shared/condaenvs
export CONDARC=${CONDA_ROOT}/.condarc
eval "$(conda shell.bash hook)"
conda activate epic-venv

Note

The first five steps must be performed every time you log into the system.

You can restore the original prompt by running conda deactivate twice.

The directory tools contains Python scripts that enable you to generate input data and analyse the output data. To use these scripts, you must make the Python interpreter aware of the modules by setting the environment variable

export PYTHONPATH=$PYTHONPATH:/work/tc066/tc066/$USER/workshop

Next: Prepare and run simulation, Previous: Accessing the Cirrus cluster, Up: Main page