Installation

The present version of the workflow requires the entire execution is carried out on the same (HPC) cluster, which must be equipped with GPUs in order to be able to run the tsunami simulations. Here is a table of the presently available clusters:

INSTITUTE CLUSTER
INGV mercalli
CINECA leonardo

The hosting institute will provide instructions on how to connect to the cluster.

Note

Before installing, check that the software and data requirements are satisfied.

Then, follow these instructions:

  1. download the package from the master branch of the gitLab repository using the "Code" button on the top-right of the repository homepage and choosing the preferred compression format;

    Warning

    The gitLab repository is private and can be accessed only by members. If you are not a member, please contact us.

  2. connect to the cluster and transfer the package to the preferred folder (hereinafter referred to as /path-to-software/);

  3. unpack the archive: a folder cheese-ptha-master will be created, whose structure is illustrated in this diagram;

  4. create/choose a working folder for IO (hereinafter referred to as /path-to-wdir/) where all the outputs will be saved;

  5. copy and rename the provided template for input files to your prefererred location, for example in your working folder (but other locations can be chosen as well): 

    cp /path-to-software/cheese-ptha-master/templates/workflow_input.json.template /path-to-wdir/workflow_input.json
cp /path-to-software/cheese-ptha-master/templates/load_env.source.template /path-to-wdir/load_env.source
    Instructions on how to fill out the JSON file are provided on the dedicated page, while the load_env.source file should be modified according to the preferred package and environment management system. We suggest Spack to build and configure the software environment, but a different package manager (e.g. conda) can be also used.

  6. unpack the initial conditions for PS seismicity (see NEAMTHM18 for details) by typing: 

    cd /path-to-software/cheese-ptha-master/inputs/SUBDUCTIONS_from_NEAMTHM18
bunzip2 INIT_COND_PS.tar.bz2
tar -xvf INIT_COND_PS.tar
    This operation can take a while and can be skipped at the beginning, but must be executed before running STEP 4 of the workflow.

In case of target sites located along the North-Eastern Atlantic coast, outside the Mediterranean Sea, the same unpacking process must be carried out for the initial conditions of sources belonging to the Mid Atlantic Ridge and to the Caribbean arc: 

cd /path-to-software/cheese-ptha-master/inputs/MidAtlantic_from_NEAMTHM18 
bunzip2 INIT_COND_PSMar.tar.bz2
tar -xvf INIT_COND_PSMar.tar

cd /path-to-software/cheese-ptha-master/inputs/Caribbean_from_NEAMTHM18 
bunzip2 INIT_COND_PSSlip.tar.bz2
tar -xvf INIT_COND_PSSlip.tar
7. create/choose a folder for the topo-bathimetric grid files (/path-to-grids/). Having this folder, and the grid files inside it, is not mandatory until the tsunami simulations are executed ( STEP 5), although the highest resolution grid can be optionally used for the selection of the Point(s) of Interest (POIs) in STEP 1.