[CPMD-list] fhi2cpmd pseudopotential converter. testers needed.

[Axel Kohlmeyer]

>>> "AS" == Ari P Seitsonen <Ari> writes:


dear ari,

thanks for taking the time to wade through messy c code.

AS> Dear Axel & other CPMD'ers,

AS>   This is a nice tool for getting the pseudos generated with the FHI98PP
AS> package to the format recognised by CPMD (an alternative is the modified
AS> 'psgen' script posted to the list earlier). Just as a comment to you
AS> points:

AS>  a) This is not so crucial, it's just used for user's convenience in a
AS>    check, it doesn't change anything in CPMD

true if you create the pseudopotentials only for your personal use.
but especially for not so experienced cpmd users, it can be 
quite confusing to get warning about inconsistent
pseudopotentials, when in fact they are not.

AS>  b) Hmm, the normalisation seems to be the same as in the 'psgen' so I
AS>    guess that you're consistent - and even correct? ;) As far as I

i did not look into the modified psgen script, only the scilab script
which does not support NLCC. the normalization factor is the same as
in the gianozzi package which i found to be consistent with the NLCC
section of the atomc (hamann) program.

AS>    remember also the wave functions need a different normalisation
AS>    [1/sqrt(4pi)] to be consistent with the other programs used to generate
AS>    pseudos for CPMD, however I believe that they are always normalised
AS>    within CPMD, so maybe it's just cosmetics (or maybe it changes
AS>    something when projecting the CPMD wave functions onto the atomic wave
AS>    functions from the pseudo file)

good idea.

that is one of the things, that are a bit annoying from time to time: 
i could not find any comments or hints, what the units are
_supposed_ to be. when comparing to potentials created with other
programs or existing pseudopotentials, i found that:

- the fhiPP wavefunctions are consistent with the ones created with
  the ps2cp converter from the gianozzi program, but not with the atomc
  code. there are also both 'versions' in the cpmd pseudopotential library.
  i just checked it and the difference is indeed a factor of 1.0/sqrt(4*pi).

- it did not make a difference when using the resulting potentials
  so i assumed that cpmd renormalizes the wavefunctions appropriately
  after reading them.

i just checked the projection to the pseudowavefunctions and it makes 
no difference there, too. if there is some authoritative information,
how the wavefunction has to be normalized, i'll change the code
accordingly nevertheless. 

best regards,
     axel.


AS>     Greetings,

AS>        apsi

AS> -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
AS>  Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
AS>  Tel +41 1 635 44 97 / Fax +41 1 635 68 38 / GSM +41 79 719 09 35      
AS>  Anschrift: Physikalisch Chemisches Institut (PCI), Universität Zürich (UniZh)
AS>  Indirizzo: Winterthurerstrasse 190, CH-8057 Zürich
AS>  Address:   Schweiz / Svizzera / Suisse / Svizra / Switzerland

AS> On Sat, 4 Sep 2004, Axel Kohlmeyer wrote:

>> 
>> hello everybody,
>> 
>> since the topic of pseudopotential creation came up in a previous
>> mail, it may be a good idea to used the attention to that topic.
>> please find attached an attempt to write a conversion tool for 
>> pseudopotentials generated with the fhi98PP package to
>> the native cpmd format. in contrast to the previously posted
>> scilab script, this is written in c (so it should be compilable
>> everywhere) and there should be no manual fixup of the 
>> resulting file needed. as far as i can tell, it seems to work 
>> correctly, but it would be very nice if any of you could review
>> and test it.
>> there are two major points, where i would like to 
>> have some proof:
>> 
>> a) the mapping of the fhi DFT functional codes to the 
>> respective CPMD code numbers.
>> b) the normalization of the NLCC density.
>> 
>> any comments are appreciated,
>> axel kohlmeyer.
>> 
>> -- 
>> 
>> =======================================================================
>> Dr. Axel Kohlmeyer                        e-mail: axel.kohlmeyer at rub.de
>> Lehrstuhl fuer Theoretische Chemie          Phone: ++49 (0)234/32-26673
>> Ruhr-Universitaet Bochum - NC 03/53         Fax:   ++49 (0)234/32-14045
>> D-44780 Bochum  http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
>> =======================================================================
>> 
>> 



--

=======================================================================
Axel Kohlmeyer       e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de
Lehrstuhl fuer Theoretische Chemie          Phone: ++49 (0)234/32-26673
Ruhr-Universitaet Bochum - NC 03/53         Fax:   ++49 (0)234/32-14045
D-44780 Bochum  http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.