[Pw_forum] Zero-norm error during PP generation

Lorenzo Paulatto paulatto at sissa.it
Sat Apr 12 13:02:19 CEST 2008


On Sab, Aprile 12, 2008 00:01, Chao Cao wrote:
>       Message from routine normalize:
>       zero norm: not a true US PP ?

Dear Chao Cao,
this error message is triggereed when self-consistency has a difficult
time to converge. If the solver cannot find a solution for a certain wfc
it sets it to zero and try with the next one, as a consequence
normalization routine willcomplain. Sometimes on the next iteration it can
find a solution for all wfcs, but the error ususally hint the presence of
some ghost state.

Yes, we will try to make the error clearer for the next release.

>  I guess I made some stupid mistake in the input file, but I simply
>  cannot find it... Any hints/suggestions? Thanks in advance.

There are not many pseudopotential libraries that have a Lanthanide
pseudopotential. There is one from N. Holzwarth, here:
http://www.wfu.edu/~natalie/papers/pwpaw/periodictable/atoms/LaN/La_GGA_input
and it has radiuses dramatically different from yours.

Actually I'm as lost as you in this matter, I can only suggest to start
with an all-electron calculation (no inputp namelist) and see what the
radiuses (average, square average, maximum) are.

I hope it helps, goodbye.

-- 
Lorenzo Paulatto
SISSA  &  DEMOCRITOS (Trieste)
+39 040 3787 511
http://people.sissa.it/~paulatto/


----------------------------------------------------------------
  SISSA Webmail https://webmail.sissa.it/
  Powered by SquirrelMail http://www.squirrelmail.org/



More information about the Pw_forum mailing list