[Q-e-developers] Bug in FFTXlib

[stefano de gironcoli]

Dear Martin,
  i modified the routine that set the fft arrays so that they are 
cleaned before being assigned.
  this avoid the error if allocate_fft is called twice (as it happens in 
the phonon code at q/=0).
  not an ideal solution but it works as before, avoiding the double 
allocation of the arrays.

  stefano

On 08/07/2016 18:02, Martin Schlipf wrote:
> There is another issue with the new structure. The logic of
> fft_dlay_set_dims is not working for linear response calculations. E.g.
> example01 of PH is broken for the si.phX.in calculation.
>
> -Martin
>
> On 08/07/16 11:45, stefano de gironcoli wrote:
>> corrected in the svn
>> thank you for reporting
>> stefano
>>
>> On 08/07/2016 10:49, Martin Schlipf wrote:
>>> Dear all,
>>>
>>> the recent changes in the arrangement of the FFT broke the code in at
>>> least one case. The do_fft_x in line 595 of FFTXlib/fft_scalar.FFTW.f90
>>> should be a do_fft_z.
>>>
>>> Kind regards
>>> Martin
>>>
>> _______________________________________________
>> Q-e-developers mailing list
>> Q-e-developers at qe-forge.org
>> http://qe-forge.org/mailman/listinfo/q-e-developers
> _______________________________________________
> Q-e-developers mailing list
> Q-e-developers at qe-forge.org
> http://qe-forge.org/mailman/listinfo/q-e-developers