The ABINIT Users Mailing List ( CLOSED )

[Igor Lukacevic <ilukacevic@fizika.unios.hr>]

Dear prof. Gonze,

after fixing the anaddb, everything works well, but, still, it seems that a new problem arises. Fixed anaddb (using unchanged mrgddb output) gives different (wrong) phonon frequencies wrt the not fixed anaddb using the edited mrgddb output file (with changed acell and rprim in it), which gives more sensible frequencies.

Also, band2eps gives a blank eps file when brav 2 is used, irrespective of the usage of fixed or not fixed anaddb, whether brav 1 option works well. Of course, the frequencies are completely different in these two cases.

If You want I could send the files I got.

Yours,

Igor Lukacevic




Xavier Gonze wrote:

163C5C3A-C487-4C40-A290-17ECC199A048@uclouvain.be" type="cite">Dear Igor,

On 04 Jan 2010, at 13:11, ilukacevic@fizika.unios.hr wrote:

Dear prof. Gonze,

thank You for Your suggestion. Anaddb works with it.

If You don't mind, could You give me a deeper explanation why this happens? Is
it a thing with a mrgddb code, which could be changed within it?

anaddb was rather restrictive.
Actually, I just improved it. Could you try whether it works with the modification below
using your non-modified DDB  ?

Best wishes,
X.

----------------------------------------------------------------
Modifications for src/98_main/anaddb.F90

(1) Declare
real(dp) :: factor

(2) Insert around line 282 (between the call to rdddb9 and the call to mkrdim) :

 if(anaddb_dtset%brav/=1)then
!  Renormalize rprim to possibly satisfy the constraint abs(rprim(1,2))=half when brav/=1
   if(abs(rprim(1,2))<tol6)then
     write(message, '(4a,i6,7a)' )ch10,&
&     ' anaddb : ERROR -',ch10,&
&     '  The input DDB value of brav is',anaddb_dtset%brav,',',ch10,&
&     '  and the one of rprim(1,2) is zero.',ch10,&
&     '  These are incompatible',ch10,&
&     '  Action : check the value of brav and rprim(1,2) in your DDB.'
     call wrtout(std_out,message,'COLL')
     call leave_new('COLL')
   endif
   factor=abs(rprim(1,2))*two
   acell(:)=acell(:)*factor
   rprim(:,:)=rprim(:,:)/factor
 endif


Recompile, and test ...

Thank You!

Igor Lukacevic