UT theoretical chemistry code forum

Hi, all,

use fvin.pl after the first run of NEB is finished, only CONTCAR is copied to POSCAR, and then these POSCARs can be used to start a second run. My question is should I also use the WAVECAR by setting ISTART=1 ( and what other output files can be used?) from first run to start the second run of NEB? Can the convergence speed improve a lot?

If you write the chgcar and/or wavecar, these get compressed and moved to the directory you specify in the vfin.pl argument. But as it is, we don't leave copies of these files for restarting. They can help speed up the first iteration, but I guess we don't find the speed savings worth the extra disk space taken up by having these large files sitting around.

Andri, what do you think? Is it worth adding a mechanism for keeping copies of the chgcar/wavecar in the running directory?

I like the setup we have now for handling WAVECAR files. If people like to use them in a subsequent run then then there is nothing simpler than coping them back from the saved directory and un-zipping them. I prefer the simplicity over adding a tag or something to decide the a copy of the WAVECAR should be left in the main directory.