Installation

This folder contains everything needed to set up your environment on the CEMDI tutorial cluster (cemdi.calculquebec.cloud).

Quick start: from a terminal on the cluster (SSH session or JupyterLab terminal), run:

bash ../setup_environment.sh

(from Tutorial/Installation/; see Tutorial/setup_environment.sh for what it does). It will:

1. Create ~/.abinit/abipy/ and copy abinit-config/default/abipy/manager.yml and abinit-config/default/abipy/scheduler.yml there -- this tells AbiPy how to launch Abinit runs on this cluster (how many cores, which queue, etc.).
2. Check that the abinit executable is on your $PATH (loading the abinit module if needed).
3. Check that the abipy Python package is importable. If you are working in JupyterLab, it already is. If you connected over plain SSH, the script creates a virtual environment and installs abipy from the cluster's local wheelhouse (pip install --no-index abipy), per the instructions from the CEMDI organizers.

Manual setup

If you'd rather do it by hand:

mkdir -p ~/.abinit/abipy
cp abinit-config/default/abipy/manager.yml ~/.abinit/abipy/
cp  abinit-config/default/abipy/scheduler.yml ~/.abinit/abipy/

Then check your installation with:

abicheck.py

About manager.yml / scheduler.yml

These are the same two files used throughout the tutorial and in AbipyExamples, just resized for the resources allocated per student on this cluster (8 CPU cores, interactive/shell execution rather than a batch queue -- see the comments in abinit/abipy/manager.yml). If the cluster turns out to require submitting through Slurm instead, that file has a pointer to the Slurm-based config used for the Rorqual cluster as a starting point for adapting it.