I have been using GULP 4.0 for a long time for geometry and phonon calculations on semiconductor nanocrystals (“clusters”). I had to modify the source code for the subroutine deffreq to normalize the eigenvectors appropriately and to print out the reduced masses I need for further calculations. I’m now trying to switch to GULP 6.0 but the comments in the new deffreq routine indicate that clusters are not handled by that routine. I can’t figure out where these calculations are done for clusters. Can anyone enlighten me?
Sorry for changing things round. The calculation of phonons for finite systems (clusters) is now computed either in the subroutine “phonon” (serial) or “phonond” (parallel) which are the same subroutines as for periodic systems of any dimensionality. When making the jump to distributed memory for parallel second derivatives it made sense to unify all the phonon calculations into a single place for each algorithm. Hope that helps.