Default configuration (system wide and user specific)

System default configuration

The system default configuration can be found in AaronTools/config.ini, and defines the built-in workflows available in AaronJr. These built-in workflows can be "included" in your AaronJr configuration file and extended as necessary. For example, one may use the workflow for TS optimizations (defined in the [Job.TS] section) and append an additional step for higher-level single-points by placing the following in their project's configuration file:

[Job]
include = TS
5 type = single-point

Note also that the CrestTS and CrestMinimum workflows already set the executable type for certain steps, so you only need to define the exec_type option for the other steps.

# snippet from AaronTools/config.ini
[Job.CrestMinimum]
1 type=optimize.changes
1 exec_type = xtb
2 type=conformers
2 exec_type=crest
3 type=optimize
4 type=frequencies

# snippet from the project's configuration file
[Job]
include = CrestMinimum
exec_type = gaussian

The snippets above will use xtb for the relaxation of changed portions of the structure (ie: substitutions and mappings applied to the template file), Crest for conformer generation, and Gaussian for everything else.

User's default configuration

You may find yourself defining the same options over and over again in your project's configuration files. If so, defaults can be set in $AARONLIB/config.ini. These settings will be automatically used, unless overridden by settings in the project's configuration file. Additionally, consider using functions to make this file more flexible.

# snippet from $AARONLIB/config.ini
[Job]
exec_type = gaussian
nodes = 1
procs = %{ $ppn * $nodes }
memory = %{ $procs * 2 }GB

Above, the default executable type is Guassian. The memory requested is calculated based on the number of processors requested. Now, we only need to set the number of cores/node (ppn) in the project's configuration.

Example user-defaults configuration

Here is the content of the file I have saved as $AARONLIB/config.ini with annotations in the form of # comments.

# remote directory for saving computational input/output files
remote_dir = /home/%{$HPC:user}/chem/%{$project}/%{$name}

[Theory]
charge = 0
multiplicity = 1
method = b3lyp
basis = !tm 6-31G(d)
        tm LANL2DZ
# options for relaxation of changes
1 method = PM6
1 basis = 
# density fitting option ignored for hybrid functionals
denfit = 1

[Job]
exec_type = gaussian
memory = %{$procs*2}GB
# some software treats memory limits as mere suggestions
exec_memory = %{$procs*16//10}GB
procs = %{$ppn*$nodes}
nodes = 1
ppn = 12
wall = 12
# options for relaxation of changes
1 ppn = 2
1 wall = 2

[HPC]
user = <username for supercluster>
host = <supercluster's host name for submitting jobs>
transfer_host = <supercluster's host name for file transfer (only needed if different from `host`)>
scratch_dir = /scratch/%{$user}
queue_type = SLURM
queue = batch