Application Guide - ADF

Introduction

ADF is a commercial computational chemistry software for first-principles electronic structure calculations.  ADF can run in parallel on the ARC clusters using MPI.

The guide shows how to

  • load the ADF module;
  • prepare a ADF job submission script and
  • submit the ADF job.

Running an ADF job

First, you need to load the module for the ADF package.  Different versions of ADF are managed through the modules; use the command: module avail adf to see what versions are available.  The sought version can be specified when loading the module, e.g.

module load adf/2014

If no version is specified, the default version is loaded.  (The default is normally the latest version available.)

Then, you need to prepare a submission script for the ADF job.  Consider the calculations to run are described in the file GO_H2O.run (this represents one of ADF's own tests); this file contains a description of the molecule (water in this case) as well as instructions as to what calculations to perform and represents the input to the ADF executable.  The submission script for this job should look something like this :

#!/bin/bash
 
#SBATCH --nodes=2
#SBATCH --ntasks-per-node=16
#SBATCH --time=01:00:00
#SBATCH --job-name=GO_H2O
 
module purge
module load adf
 
. $ADFHOME/adf.sh
run-adf GO_H2O.run

The script requests a two cluster nodes (nodes=2), each running 16 MPI processes (ppn=16 or --ntasks-per-node=16).  With ADF, this is the only mechanism to control parallelism, and the information contained in that line directly determines the number of processes during the run.  The number of MPI processes per node should always be 16 to fill the node.

The script adf.sh which sourced in the submission script prepares the parallel environment for ADF and defines a run directory with a name that is unique to each job.

The final line of the submission script start ADF in the run directory, taking GO_H2O.run as input.

Supposing script listed above is called GO_H2O.sh, the jobs is sent to the execution queue with the following command:

 

sbatch GO_H2O.sh

 

Example Job

We have provided an example job environment (based on the scripts above) which you can use to familiarise yourself with the batch system or modify for your own jobs. To extract it:

 

cd $DATA tar xf /system/software/examples/scripts/adf.tgz

 

This will create a directory named adf-example, to run the example job:

cd adf-example
sbatch adf_slurm.sh

 

Service area: 

Written by IT Services. Latest revision 28 March 2017