In this exercise you will perform a series of basic convergence tests with respect to several input parameters.
When running DFT calculations, several approximations are used in order to be able to run the code on a computer.
In order to determine whether the accuracy of the desired results is being affected by these approximations, a series of convergence tests has to be performed.
The basic idea behind them is to vary a parameter (one at a time!) inherent to the accuracy of our calculation and observe how certain quantities vary with it. When these variations are below a threshold that represent the desired output accuracy, we say that our calculation is "converged" with respect to that parameter.
You might rightfully ask: "Why compromise at all and not use extremely high values for these parameters?". The answer to this is COMPUTATIONAL COST (i.e. time and amount of computational resources that are needed to obtain the final result).
You will notice while doing the exercises that a higher value of a convergence parameter will also result in a longer calculation time.
Computational resources have a cost and most importantly are usually limited - the goal is to reduce the computational cost while still obtaining physically relevant results.
For more details on each approximation, refer to the documentation of each subsection and (the latest version of) the file "handout.pdf" provided for this workshop.
a. Convergence of total (absolute) energies with respect to cutoff energies
b. Convergence of forces with respect to cutoff energies
c. Convergence of the total (absolute) energies with respect to the size of the k-points mesh
d. Convergence of forces with respect to the size of the k-points mesh
e. Convergence of the total energy differences with respect to energy cutoff