I am a beginner in Quantum Espresso and I am trying to do geometrical optimization using Quantum Espresso. Instead of using calculation = 'relax', I wish to use my own optimizer by using calculation = 'scf' and then do iterations. It is working fine but I wish to reduce the computation time. Therefore, I was trying to use the energy or wavefunction obtained from my last scf calculation as the initial guess for my next scf calculation because I am moving the atoms of my system infinitesimally. The problem I am facing is that I can only give the coordinates of my atoms as input and then QE does its own calculations, therefore, I do not get the opportunity to provide a good initial guess to it (which would be energy/wavefunction from my last calculation). Is there a way to provide the initial guess to QE? A way in which it could work can be to set *wf_collect = .TRUE.* (which collects the wavefunction in the tmp folder) and then setting restart mode = 'restart' for next calculation which I think would use the wavefunctions from last calculation but I am not able to see clearly how these things can be executed? Also, what should be force and energy tolerance values for a Si system for convergence of a geometrical optimization problem to reduce the number of iterations?
Thanks Mitul Mundra -------------- next part -------------- An HTML attachment was scrubbed... URL: http://pwscf.org/pipermail/pw_forum/attachments/20130928/51e124fa/attachment.html
