[Wien] lapw2 error

[tkamiya at rlem.titech.ac.jp]

Dear Dr. Blaha,

  Thank you for the answer and sorry for bothering you with
the silly question.

  Finally, I could get through the calculation. The way was
just that you said, I re-started it from the StructGen stage
in an empty directory on single execution mode. And I should also
say that my probelme is nothing to do with the Dr. Ma's case.

  I would be mixed up different problems. I had actually done
the same process, i.e. from the StructGen stage for a few times.
In those cases, I first performed parallel execution for Non-SO 
case. When they were converged, I continued them with SO interaction,
however, they are crashed in lapwso stages (I may have to check 
my parallel system). Then I changed to single execution.
The calculations passed lapwso but lapw2 was stopped
with the inconsistent k-point error. I'm afraid that I might did 
something wrong when changed to single execution mode.

  This problem does not appear if I do not use parallel execution.
I am again sorry that I can not survey the problem in deeper because
the calculation takes long time. As for the short calculation time
seen in my previous mail, that is because I had reduced k-points 
to 1 k-point as I did not want to waste time to seek a way 
to solve the problem. La2O2CdSe2 has 14 atoms in an unit
cell of P42mc, so it takes almost whole day if I use 1000 k-points.


>Hi,
>
>Now I'll say it similar to Kevin: With so little information it is very
>difficult to help. When you say "sometimes it runs sometimes" you cannot
>expect a reasonable help.
(cut)
>>   For some parameters (I am not sure), several SCF cycles
>> went well, but after then lapwso stopped or lapw2 stoped
>> with the FERMI error concerning inconsistent k-points.
>> For some parameters, first SCF cycle stops with the message
>> that I reported.
>
>I do not believe it at all! During a scf calculation this message cannot
>occur unless you change the k-mesh,....
(cut)

Sincerely yours,
Toshio KAMIYA