Final SCF calculation and Pulay stress.

[Aliaksandr Yakutovich]

Dear all,

We are trying to implement a workflow that performs cell optimisation. An important part of the workflows is the Pulay stress [1] analysis.- estimation of the error in the stress tensor introduced by the incompleteness of the basis set. 

Essentially, to compute the Pulay stress one needs to re-run the SCF (re-initialising the plane-wave basis set) with the final geometry.

At the same time, I believe, re-initialisation of the basis set happens before the very final SCF run (the one that starts with "Reevaluating energy at the minimum"). Can someone confirm that this is correct?

If that is correct, is there a possibility to print the stress tensor at this very final SCF step? We didn't find an option to enable that. Having this information printed would allow us to avoid re-running the SCF just to compute the final stress (the Pulay stress).

Thanks,

Best regards,

Sasha

[1] https://iopscience.iop.org/article/10.1088/0953-8984/2/19/007

[Aliaksandr Yakutovich]

Dear all,

If I understand correctly, restarting the CP2K calculation would not change anything in that regard.

This is because (unlike "conventional" plane wave codes) CP2K anyways recomputes the basis for every SCF loop.

It happens as the step of the GPW algorithm when mapping the Gaussian basis to the plane waves.

Best,

Sasha