[Wannier] Generating .mmn matrix for all pais of k-points

[Dominik Gresch]

Dear Vahid,

It seems to me the mp_grid is not actually used when running with -pp, 
but I'm not sure about that. You can try just setting mp_grid to 
'num_kpoints 1 1', and see what happens, but for now I'd be careful with 
the results. As a quick check maybe you can set mp_grid to some 
ridiculous value and check if the results are still the same.

If it's actually the case that the mp_grid variable is not used in 
postproc_setup mode, then we should change the code such that it doesn't 
raise an error in this case.

Best,

Dominik


On 17.03.2017 15:29, Vahid Askarpour wrote:
> Dear Dominik,
>
> I installed qe-6.1, which uses wannier90 version 2.1. I ran a 
> self-consistent run followed by a non self-consistent run on the 
> following k-points:
>
>    0.0000000   0.0000000   0.0000000  0.0312500
>    0.0000000   0.0000000   0.2500000  0.2500000
>    0.0000000   0.0000000  -0.5000000  0.1250000
>    0.0000000   0.2500000   0.2500000  0.1875000
>    0.0000000   0.2500000  -0.5000000  0.7500000
>    0.0000000   0.2500000  -0.2500000  0.3750000
>    0.0000000  -0.5000000  -0.5000000  0.0937500
>    0.2500000  -0.5000000  -0.2500000  0.1875000
>
> Then I used the following si.sa.win file:
>
> begin projections
>  Si:sp3
> end projections
> num_bands 12
> num_wann   8
> iprint   2
> dis_win_min -1000.000
> dis_win_max    17.000
> dis_froz_min -1000.000
> dis_froz_max     6.500
> num_iter     500
> dis_num_iter 1000
>
> nnkpts 8
> begin nnkpts
> 1 1 0 0 0
> 1 2 0 0 0
> 1 3 0 0 0
> 1 4 0 0 0
> 1 5 0 0 0
> 1 6 0 0 0
> 1 7 0 0 0
> 1 8 0 0 0
> end nnkpts
>
> begin unit_cell_cart
> bohr
> -5.084373 0 5.084373
> 0 5.084373 5.084373
> -5.084373 5.084373 0
> end_unit_cell_cart
>
> begin atoms_frac
> Si       0.000000000   0.000000000   0.000000000
> Si       0.250000000   0.250000000   0.250000000
> end atoms_frac
>
> begin kpoints
>    0.0000000   0.0000000   0.0000000
>    0.0000000   0.0000000   0.2500000
>    0.0000000   0.0000000  -0.5000000
>    0.0000000   0.2500000   0.2500000
>    0.0000000   0.2500000  -0.5000000
>    0.0000000   0.2500000  -0.2500000
>    0.0000000  -0.5000000  -0.5000000
>    0.2500000  -0.5000000  -0.2500000
> end kpoints
>
> Here I used the same k-points as before and asked for nearest 
> neighbors for k-point=1
>
> I ran wannier90.x -pp si.sa in parallel and get the following error:
>
> Error: You must specify dimensions of the Monkhorst-Pack grid by 
> setting mp_grid
>
> How do you specify an mp_grid if the k-points are chosen randomly? Or 
> am I missing something in the input?
>
> Thanks,
>
> Vahid
>
>
>> On Mar 16, 2017, at 8:46 PM, Vahid Askarpour <vh261281 at dal.ca 
>> <mailto:vh261281 at dal.ca>> wrote:
>>
>> Dear Dominik,
>>
>> Thank you so much for your quick response. I will follow your 
>> suggestions.
>>
>> Best wishes,
>>
>> Vahid
>>
>>> On Mar 16, 2017, at 8:35 PM, Dominik Gresch <greschd at phys.ethz.ch 
>>> <mailto:greschd at phys.ethz.ch>> wrote:
>>>
>>> Dear Vahid,
>>>
>>> You can use the 'nnkpts' parameter (new in version 2.1) to 
>>> explicitly specify for which k-point pairs the overlap should be 
>>> calculated. This just side-steps the way that Wannier90 tries to 
>>> calculate the nearest neighbours, and prints the neighbours you 
>>> supply directly to the .nnkp file. To get the .mmn output for all 
>>> pairs of k-points, you could generate the nnkpts block such that it 
>>> includes all the overlaps you need.
>>>
>>> For the interface to QE, you need to use the 'regular_mesh = 
>>> .false.' when using pw2wannier90 to allow for non-regular k-grids to 
>>> be used (new in QE v.6).
>>>
>>> Best regards,
>>>
>>> Dominik Gresch
>>>
>>>
>>> On 17.03.2017 00:13, Vahid Askarpour wrote:
>>>> Dear Wannier Community,
>>>>
>>>> I am attempting to include impurity scattering rate calculation in 
>>>> the EPW calculation. At present, EPW determines the scattering 
>>>> rates only due to electron-phonon scattering.
>>>>
>>>> To include impurity scattering, one needs the overlap intergal for 
>>>> the lattice-periodic part of the wave function |u_nk> between any 
>>>> two k-points. The .mmn files creates such a file but it is only for 
>>>> the nearest neighbor k-points. Before attempting to redo a separate 
>>>> code, I have two questions:
>>>>
>>>> 1. Is it possible to output the .mmn file for all pairs of k-points?
>>>>
>>>> 2. Is it possible to do 1 not for a k-grid but for a list of random 
>>>> k-points? I suspect this can be done using
>>>>
>>>> begin kpoints
>>>> # of k points
>>>> …
>>>> …
>>>> end kpoints
>>>>
>>>> and the same k-points must be used in the preliminary 
>>>> self-consistent and non self-consistent calculations with QE.
>>>>
>>>> I ask question 2 because for thermoelectric properties, one needs 
>>>> to select k-points with energies  ~0.5 eV above the CBM. So it is 
>>>> possible to generate a very dense k-grid (100^3) and only include 
>>>> the selected k-points in the impurity scattering rate calculation.
>>>>
>>>> I would appreciate your help with the above questions.
>>>>
>>>> Vahid
>>>>
>>>> Vahid Askarpour
>>>> Department of Physics and Atmospheric Science
>>>> Dalhousie University,
>>>> Halifax, NS, Canada
>>>>
>>>>
>>>> _______________________________________________
>>>> Wannier mailing list
>>>> Wannier at quantum-espresso.org
>>>> http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier
>>>
>>
>
>
>
> _______________________________________________
> Wannier mailing list
> Wannier at quantum-espresso.org
> http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/wannier/attachments/20170317/99b10a06/attachment.html>