Introduction¶

Welcome to the AiiDA ChemShell plugin documentation. This plugin provides an AiiDA workflow plugin for the ChemShell multiscale computational chemistry package developed at STFC.

Quick Start¶

This plugin should be installed alongside the aiida-core package, typically using the pip python package manager. The latest release is available from PyPI:

pip install aiida-chemshell

Alternatively, to install the current development version from a local checkout of the source:

cd aiida-chemshell
pip install .

Requirements¶

To use this plugin a configured AiiDA profile and computer are required, as well as a compiled ChemShell executable. For more information on how to set up AiiDA refer to the AiiDA documentation.

Setup¶

To configure the ChemShell plugin you will need to create a new AiiDA code object. The following is an example of a basis YAML configuration file for a ChemShell code:

label: chemshell
description: ChemShell
computer: localhost
filepath_executable: chemsh.x
default_calc_job_plugin: chemshell
use_double_quotes: false
with_mpi: true
prepend_text: ""
append_text: ""

Write this to a file named chemshell.yml ensuring the value for computer matches the label of your computer configured in the previous step. The code can then be configured by running:

verdi code create core.code.installed --config chemshell.yml -n

If successful, this will create a new AiiDA code with the label chemshell which can then be used to run ChemShell calculations within the AiiDA framework.

ChemShell with MPI¶

The configuration of an MPI enabled ChemShell code differs slightly depending on whether the code instance uses the chemsh wrapper script or the chemsh.x executable directly:

If you have a parallel (MPI enabled) build of ChemShell but are using the chemsh wrapper executable, set with_mpi: false. The underlying plugin will handle how to invoke the MPI runtime.
If you are calling the chemsh.x executable directly, then with_mpi: true is required to run parallel jobs.

The number of MPI processes is taken from the metadata.options.resources of the calculation, defaulting to one machine with four processes per machine (see Inputs, Parameters and Outputs).