The ABINIT Users Mailing List ( CLOSED )

[Matthieu Verstraete <mjv500@york.ac.uk>]

Hello Mavrin,

please always post question to the forum. I am only one of 3 or 4 people 
who could answer your question, and there are (modesty aside) probably 
dozens of other who could benefit from the answer.

The answer: you removed the shiftk = 0.0 0.0 0.0 lines in the input.
By default abinit choses 0.5 0.5 0.5, but this is problematic in elphon 
as it is coded right now: your kpoint mesh (not just qpoint mesh) must go 
through the gamma point. I think this is mentioned in the tutorial text.

In some cases this restriction could be lifted, but in your case it would 
not be this simple: the important point is that the kpoint mesh must be 
compatible with symmetry. In your fcc case you need the 4 shifts used in 
the tutorials as well. Using only one (1/2 1/2 /2) gives you a grid which 
does not posess the full symmetries of the lattice. That is why anaddb is 
complaining: it tries to complete the kpoints using symmetries, but 
symop*k = k' which is not in the GS grid. Bad.


Matthieu

--
================================================================
Dr. Matthieu Verstraete                 mailto:mjv500@york.ac.uk
Dept. of Physics, University of York,     tel: +44 1904 43 22 08
Heslington, YO10 5DD York, United Kingdom fax: +44 1904 43 22 14

---------- Forwarded message ----------
Date: Fri, 9 Mar 2007 16:18:31 +0300
From: Mavrin <mavrin@isan.troitsk.ru>
To: verstraete@pcpm.ucl.ac.be
Subject: Help

Dear Dr. Matthieu Verstraete,

 I have a question on the ABINIT EPI calculations.
 I use intel_DOSWin_seq_4.6.5 code. The EPI calculations for
 Aluminium, as it is done in ABINIT tutorial, were successfull for me.
 I tried to apply this code for tin with the FCC structure
 and natom=2 in unit cell. I have employed, as examples of input
 files, the files from the ABINIT tutorial with a correction of natom
 and 50sn.pspnc pseudopotential from the ABINIT site (my telphon_1.in
 file in the attachement). Input data (grids and others) are not
 optimal yet.

 After the run of mrgddb-code with 18 _DDB files I have seen the final
 code: "+mrgddb: the run completed successfully".

 After the run of mrggkk-code with 1 _WFK file+18 _1WF files I have
 seen the echo:"normal input for 1WF file".

 But the run of anaddb-code was not successfull. In log-file one can
 find the echo "the interatomic forces have been obtained". And
 somewhat below: "read_gkk error: FSkpt 65 is not the symmetric of one
 of those found in the GKK file".

 Give advice how to correct the input files for the successfull run.


--
Best regards,

Dr. Boris Mavrin                          mailto:mavrin@isan.troitsk.ru
Institute of Spectroscopy of RAS
Troitsk Moscow reg. 142190
Russia

Attachment: telphon_1.in
Description: Binary data