[Pw_forum] Restarting a vc-relax calculation

[Joshua Davis]

Dear QE users

I am calculating with QE 5.1, and I did a vc-relax calculation on my system
and found the difference between energy of 2nd to last step and the last
step to be pretty small as it was only a 0.5 mRy difference.  (This is were
the G-vectors of the initial and final structure are compared.)  The energy
difference seems small enough to suggest my cut-offs are fine, but I am new
to QE and based on what I know about other codes and how people relax a
cells with the other codes, is to 'restart' the calculation and to relax
with the new defined G-vectors of the cell.

However, when I set restart_mode = 'restart' the calculation started from
the new structure, but basically ran relax instead of vc-relax. So for the
second calculation the cell was not relaxed.

My questions are:

Is this a bug? (I'm sure you don't like new users asking about a bug)

For those who run vc-relax a lot do you just run it once? and if the energy
between 2nd to last and the last structure are too different do you just
punch up the wfcecut and start the calculation from the very beginning? I
had already done a convergence test on ecut and kpoints.

Are my results fine and I should just move on to what I want to do next
with the structure?  My cell only has 14 atoms.


Thanks for the help
----------------------------------------------------------------------------------------------------------------
Joshua D. Davis

Graduate Assistant
Department of Chemistry
Michigan State University

East Lansing, MI 48824
-----------------------------------------------------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20141113/e043d913/attachment.html>