Automation Tips

Awk

Awk is best used when dealing with files of data that is arranged into columns and rows. For example, if you want to just grab the first column of a file you would type awk '{print $1}' filename. If you want to convert the data from a column vetor to a row vector for copy/paste purposes you can use echo $(awk '{print $1}' filename).

Batch Jobs

There are occasions where running a bunch of similar jobs with small changes can be useful. An example is when you are trying to converge the number of k-points or the energy cutoff. One way of doing these batch jobs is to execute the following commands on the command line.

    for i in $(seq start step stop); do
    cp -r files/ $i
    cd $i
    sed -i "s/encut/$i/g" INCAR
    sed -i "s/encut/$i/g" vasp*run*
    sbatch vasp*run*
    cd ..
    sleep .1
    done 

These commands create a new folder named i and look for the phrase encut within the INCAR file—to set the ENCUT parameter—and the sbatch file—to name the calculation—and replace it with the value i. It then submits the job. This is done for i starting at start with a step size of step until stop is reached. Be careful to use double quotes and not single quotes in the lines starting with sed. Usually it does not matter, but because we are using the variable i we need to use double quotes.