[QE-users] GWL FFT questions

Eric Glen Suter esuter at uga.edu
Mon Apr 23 15:19:10 CEST 2018


Hello all,


I've been trying to run a GW calculation using the GWL package. I took a sample Si run from here: http://www.gwl-code.org/tutorial_gwl.pdf


However, on the step that involves executing pw4gww.x I get the error:


 Before initialize_fft_custom   15.000000000000000        1.0000000000000000

          15          15          15

          15          15          15


     Planes per process (custom) : nr3t =  15 nr3p =   15 ncplane =   225


     Proc/  planes cols     G


        1    15    121     1021


 GRID G to R          15          15          15

 GRID G to R         511

  *            1  *          121  /=          225


 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

     Error in routine  fft_type_set (7):

      inconsistent number of sticks

 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%


     stopping ...


>From the output file I've managed to piece together that this error happens in the routine setup_exx_cus, specifically as it passes ex_cuss%fft_r2g to initialize_fft_custom. I noticed that this routine gets invoked twice, once for exx_cus%fft_g2r and once for ex_cuss%fft_r2g. But in the initial call of setup_exx_cus, it passes a value for cutoff of 40.d0 Ryd that is then used by initialize_fft_custom for the fft_r2g structure. Is there a reason this cutoff is hard-coded as 40? And if not, could it be the reason fft_type_set is getting the wrong number of sticks?


I'm not sure if there's a specific forum for GWL that I should be posting this in, but any help or guidance would be greatly appreciated. Thank you in advance for your time.


Best regards,

Eric Suter

University of Georgia

Department of Physics and Astronomy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20180423/f79df9bd/attachment.html>


More information about the users mailing list