My Community

Hiya,

I am performing many noncollinear calculations in which both the ions and the magnetization density relax simultaneously. Unfortunately the jobs are often timed out by the local system.

Is it possible to begin a continuation job when this happens? I know this is possible for a collinear job using CONTCAR, but I think noncollinear jobs are different because one also needs to begin again with the most recent state of magnetization.

Any other ideas on how to get around this problem would also be much appreciated.

Thanks

Ben

Make the jobs shorter (fewer NSW steps) and continue from the last CONTCAR using also the CHGCAR and the WAVECAR written (ISTART and ICHARG = 1), then you have also the last magnetic configuration considered.