[gpaw-users] Wigner-Seitz LDOS problem

[Toma Susi]

Hi,


All right, the Wigner-Seitz LDOS works in the grid mode and gives reasonable results, so I can corroborate Miguel's modifications. However, one complex/float issue still remains:

/home/lv70811/susit5/gpaw_inst_trunk/lib64/python2.6/site-packages/gpaw/utilities/dos.py:291: ComplexWarning: Casting complex values to real discards the imaginary part
  + np.dot(Delta_p, P_p))

So a thorough update should probably also address the (assumed) complexity of P_p.

Cheers,
Toma



> On 27 Jun 2016, at 12:20, Toma Susi <toma.susi at univie.ac.at> wrote:
> 
> Ah, explanation for the 3514: I was using the plane-wave mode and that is their number, will check if a grid calculation works.
> 
> 
>> On 27 Jun 2016, at 12:03, Toma Susi <toma.susi at univie.ac.at <mailto:toma.susi at univie.ac.at>> wrote:
>> 
>> Dear all,
>> 
>> Is the Wigner-Seitz LDOS projection code still broken? Miguel Caro reported an issue already in 2013, and it appears to me this was not addressed:
>> https://listserv.fysik.dtu.dk/pipermail/gpaw-users/2013-October/002434.html <https://listserv.fysik.dtu.dk/pipermail/gpaw-users/2013-October/002434.html>
>> 
>> However, just commenting out the assert statement and modifying the form of P_p in dos.py does not seem to lead to a working projection for me, leading to a broadcast error:
>> 
>> Traceback (most recent call last):
>>   File "pdos.py", line 7, in <module>
>>     energySi, pdosSi = calc.get_wigner_seitz_ldos(1, spin=0, npts=201, width=0.1)
>>   File "/usr/local/lib/python2.7/site-packages/gpaw/aseinterface.py", line 361, in get_wigner_seitz_ldos
>>     energies, weights = raw_wignerseitz_LDOS(self, a, spin)
>>   File "/usr/local/lib/python2.7/site-packages/gpaw/utilities/dos.py", line 298, in raw_wignerseitz_LDOS
>>     np.where(atom_index == a, psit_G, 0.0))**2)
>> ValueError: operands could not be broadcast together with shapes (16,16,63) (3514,) () 
>> 
>> where (16,16,63) is my gpts mesh, not sure what the 3514 is about (there are 2025 irreducible k-points and the calculation has spin-polarization and unoccupied states).
>> 
>> Any ideas?
>> 
>> Dr. Toma Susi
>> Principal investigator
>> FWF Project HeQuCoG
>> University of Vienna, Austria
>> http://mostlyphysics.net <http://mostlyphysics.net/>
>> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/gpaw-users/attachments/20160627/373e6d79/attachment-0001.html>