Inpgen2 fails sometimes to generate a kpoint mesh for 'primitive tetragonal' systems after shifting the crystal

Summary

inpgen2 fails to generate proper k-mesh, a JuDFT-WARNING is thrown. (Occurrence rate: 0.1% of a structure set) I am not sure what problems this will cause for Fleur.

Input and output file:

Here 3 examples of where this happens:

1.aiida.in out

2.aiida.in out

3.aiida.in out

Compute environment

inpgen2, develop version on iff pc.

This is BUG because:

inpgen2 should succeed in generating a proper k-mesh for these systems

The problem only occurs if:

So far only for 'primitive tetragonal' systems and after the crystal was shifted. Therefore, it might well be that the symmetry of the system is not really tetragonal and the issue arises from the shifting.

Console output and other logs

tail out:

 primitive       tetragonal
 WARNING!!!!!!!
nmop-Parameters not in accordance with symmetry
    2    1
 we have set nmop(2) = nmop(1)
    6    6    6 new val for nmop:
 ***************ERROR***************
 k point mesh not compatible with symmetry (1)
 ***************ERROR***************

stderr:

************juDFT-Warning*****************
Error message: k point mesh not compatible with symmetry (1)
Error occurred in subroutine: kptmop
*****************************************
Note: The following floating-point exceptions are signalling: IEEE_DENORMAL
STOP juDFT-STOPPED

stdout:

      Welcome to FLEUR - inpgen   (www.flapw.de)
      Development Version     (based on MaX4.0)
 Found inversion center at finite position.
 Shifting crystal by:
  -0.2500000000  -0.2500000000   0.0000000000

 Processing k-point string
 Generating default k-point set
           6           2           6          72

Edited Sep 08, 2020 by Jens Bröder