The ABINIT Users Mailing List ( CLOSED )

[lorente@irsamc.ups-tlse.fr]

Hi,

We are running fairly long jobs with spin polarisation.

We have tried unsuccessfully to restart the calculation from a wavefunction 
written to the disk. Apparently ABINIP reads the wavefunction but the 
convergency considerably worsens! It is as if there were a contradiction 
between the read wavefunction and the rest of the parameters despite the fact 
that we keep them the same. We are not moving the ions. We try to only 
converge in the SCF cycles. The restart file contains then the same 
parameters as the initial file PLUS irdwfk set to 1.

We are using abinit version 4.0.3 on an SGI machine which otherwise runs 
fine. 

Do we need to read in other files? Do you recommend to use restartxf even if 
we do not use Broyden minimisation?

Thanks a lot,

Nicolas Lorente