The ABINIT Users Mailing List ( CLOSED )

["zhangtingPKU" <zhangting@pku.edu.cn>]

Michel C�t椋�您好�?/FONT>

 

　　Thanks very much!

    The q-vector grid is generated with ABINIT itself, with 1x1x16, kptopt=1. If I use 1x1x8 to generate q-vector, I can get the (0 0 0.0625), but the density of q-vector grid is therefore sparse. If I want to mantain the density of q-grid, should I use such a q-grid:

        0 0  0

        0 0 0.0625

        0 0  0.125

                   0  0   0.1875

                  0   0   0.25

                  0   0   0.3125

                  0   0   0.375

                  0  0   0.4375

                  0  0   0.5

         Like use the 1x1x16, shiftk= 0  0  0. But such a q-vector setting will cause other problems in anaddb code. Must I use a thin q-grid ?

         

 

======== 2006-07-17 15:18:00 您在来信中写道： ========

 

Having looked at your input file, it seems that the problem is in the q-vector you are asking.

When you are using only one set of wave functions on a grid (in your case 1 1 16 shifted by 0 0 0.5) you can get the phonons corresponding to difference in the k-points grid, meaning the non-shifted grid 1 1 16. Hence, you do not have access to q-vector 0 0 0.03125, only the q-vector 0 0 1/16=0.0625.

I hope this help,

Michel


Le 14/07/06 06:36, «?zhangtingPKU?» <zhangting@pku.edu.cn> a écrit?:

Dear ABINIT users:

   I'm doing phonon calculation with ABINIT code. Firstly I did a scf calculation to generate the ground state wavefunction, whose input file is the scf.in attached. Next, I used this WFK file to perform phonon calculation(the input file is the Q_3125E-2.in, which calculate the phonon at Q-point (0 0 0.03125)), but mistake appears at the end of the log file(in the log-Q_3125E-2 attached):
----------------------------------------------------------------------------------------------------------       hdr_check: WARNING -
  Restart of self-consistent calculation need translated wavefunctions.
  Indeed, critical differences between current calculation and
  restart file have been detected in:
        * the number, position, or weight of k-points
        * the format of wavefunctions (istwfk)
================================================================================

 inwffil: ERROR
  The file Q_3125E-2o_DS1_WFK cannot be used to start the
  present calculation. It was asked that the wavefunctions be accurate, but
  at least one of the k points could not be generated from a symmetrical one.
  dksqmax=    4.560066E-05
  Action: check your wf file and k point input variables.

 leave_new : decision taken to exit ...
----------------------------------------------------------------------------------------------------------

      I've never met such a problem before.With the same method,graphite's phonon can be calculated with no error. What's the meaning of this? Why the WFK file can't be used by phonon calculation? I've checked the k-point in the two input files, but haven't found problem. Thanks!


　　　　　　　　致
礼！


　　　　　　　　zhangtingPKU
　　　　　　　　zhangting@pku.edu.cn
　　　　　　　　　　2006-07-14

--
***************************************************************
  Michel Cote                          tel: +1 (514) 343-5628     
  Professeur agrégé                   fax: +1 (514) 343-2071               
  Département de physique
  Université de Montréal
  C.P. 6128, succ. Centre-ville
  Montréal (Québec) H3C 3J7          Michel.Cote@umontreal.ca
  Canada                   www.phys.umontreal.ca/~michel_cote
***************************************************************

= = = = = = = = = = = = = = = = = = = = = =

　　　　　　　　致
礼！

 

　　　　　　　　　　　　　　zhangtingPKU

　　　　　　　　　　　　　　zhangting@pku.edu.cn

　　　　　　　　　　　　　　　　　2006-07-18