The ABINIT Users Mailing List ( CLOSED )

[<6671011@163.com>]

Dear friends,


I am doing a time-consuming linear response-function calculation---now at the
phonon+Efield stage; the problem is that our cluster, which does not support
hibernation, is going to shut down. I hope to reuse the present results in the
future.


Is it possible to realize the above mission by controlling the parameter
rfatpol? For example, if I want to do the perturbation of n atoms, there are
two possible ways: 

(1) do the calculation in a single DATASET by using 
      rfatpol   1   n

(2) do the calculation in two separate calculations (then merge their DDBs)

     rfatpol   1   i  
     rfatpol    i    n


I tried the above routines by testing the built-in test-file, tnlo_2. However,
they do not give the same result.  Could you give me some suggestions?



Thanks.

Sincerely,
Gunagfu Luo