[Wien] Error in BerryPI Ver. Mar 15, 2021

Mon Oct 4 16:56:38 CEST 2021

Dear Vitalii,

thanks again for reporting the issues. It pointed to some limitations in parsing the struct file related to writing the oxidation state along the side of the element name (e.g., Fe3+). It is OK from WIEN2k perspective, but BerryPI was not ready for that change. Also, I fixed reading of the angles (20 instead of 120 ged).

I suggest updating BerryPI (please see instructions here https://github.com/spichardo/BerryPI/wiki/Installation)

I encountered another issue trying to run your structure. The calculation yields a metal. Here are the band occupancies:

       TEMP.-SMEARING WITH    0.00200 Ry
          -S / Kb           =  -7.03015466
          -(T*S)/2          =  -0.00351508
          Chem Pot          =   0.47784960
         Bandranges (emin - emax) and occupancy:
:BAN00276: 276    0.455493    0.471354  0.98266438
:BAN00277: 277    0.461954    0.473011  0.96308654
:BAN00278: 278    0.467157    0.473773  0.95117094
:BAN00279: 279    0.469344    0.476153  0.87986424
:BAN00280: 280    0.472317    0.476920  0.82870047
:BAN00281: 281    0.474386    0.479084  0.64359769
:BAN00282: 282    0.478369    0.480901  0.30835266
:BAN00283: 283    0.479605    0.481091  0.21542985
:BAN00284: 284    0.480418    0.482618  0.15653925
:BAN00285: 285    0.480901    0.482824  0.13158760
:BAN00286: 286    0.481049    0.482827  0.10831523
:BAN00287: 287    0.483618    0.484738  0.03915318
:BAN00288: 288    0.483618    0.485884  0.02039005
:BAN00289: 289    0.485927    0.489269  0.01128127
:BAN00290: 290    0.486318    0.490368  0.00721195
:BAN00291: 291    0.487431    0.490754  0.00403120

You use spin polarization (FM, AFM?). My setup was ferromagnetic. Maybe there is an AFM setup with the band gap?

My BerryPI run stops after LAPW2 in getNumberOfBands, which is logical since we cannot calculate the polarization of a metal.

File "/home/rubel/WIEN2k_21.1b/SRC_BerryPI/BerryPI/calculations.py", line 135, in getNumberOfBands
    iHOMO = theList[-1][0] # highest occupied band index
IndexError: list index out of range

Also, I would consider setting the charge convergence (e.g., -cc 0.001) in addition to the energy convergence. Otherwise the calculation may be not well converged. In addition, your cell is quite big (-p option can be useful in WIEN2k and BerryPI).

I hope it will help
Oleg

________________________________________
From: Wien <wien-bounces at zeus.theochem.tuwien.ac.at> on behalf of turchenko <turchenko at jinr.ru>
Sent: Saturday, October 2, 2021 10:15
To: wien at zeus.theochem.tuwien.ac.at
Subject: [Wien] Error in BerryPI   Ver. Mar 15, 2021

Dear WIEN2k Creators and Users,
I am running Wien2k v.21.1 on a machine
Z370/i7-8700K/32G2133/256G_M.2_PCIE/T2TB_W/8G_GTX1070/600W/ with
operating system Centos 7, fortran compiler INTEL
(parallel_studio_xe_2020), and Python 2.7.5
The purpose of my calculations is to get quantity of spontaneous
polarization in hexagonal cell of barium ferrites therefore I try to use
software BerryPI.
1) For initialization of calculation I have used follow parameters: XC
13 ; ecut -6.8 ; RKmax= 7 ; Gmax= 12 ; kgen= 25 ;

2) I am running from command line               runsp_lapw -ec 0.001

3) I am running from command line               berrypi -sp -k 2 2 2

This is the bottom part of the output file:
1st_BerryPI_SG186_z0_PBE.outputberry
…
[ BerryPI ] Calling command: x w2w -dn
[ BerryPI ] Finished Berry phase computation for
1st_BerryPI_SG186_z0_PBE
[ BerryPI ] Performing main calculation on 1st_BerryPI_SG186_z0_PBE
Reading case.struct file
   Lattice type: H
   Number of inequivalent atoms: 18
   a = 11.109794 bohr
   b = 11.109794 bohr
   c = 44.089276 bohr
   alpha = 90.0 deg
   beta  = 90.0 deg
   gamma = 20.0 deg
   Lattice type: H
   Reciprocal conventional lattice vectors br1_rec (2pi/bohr):
     [ 6.530452e-01,  0.000000e+00,  0.000000e+00]
     [ 3.265226e-01,  5.655537e-01,  0.000000e+00]
     [ 0.000000e+00,  0.000000e+00,  1.425105e-01]
   Reciprocal primitive lattice vectors br2_rec (2pi/bohr):
     [ 6.530452e-01,  0.000000e+00,  0.000000e+00]
     [ 3.265226e-01,  5.655537e-01,  0.000000e+00]
     [ 0.000000e+00,  0.000000e+00,  1.425105e-01]
   Real space conventional lattice vectors br1_dir (bohr):
     [ 9.621364e+00, -5.554897e+00,  0.000000e+00]
     [ 0.000000e+00,  1.110979e+01,  0.000000e+00]
     [ 0.000000e+00,  0.000000e+00,  4.408928e+01]
   Real space primitive lattice vectors br2_dir (bohr):
     [ 9.621364e+00, -5.554897e+00,  0.000000e+00]
     [ 0.000000e+00,  1.110979e+01,  0.000000e+00]
     [ 0.000000e+00,  0.000000e+00,  4.408928e+01]
   Unit cell volume: 4712.76312293 bohr3
Error in: {'Z-Coord': [0.0, 0.5], 'Y-Coord': [0.0, 0.0], 'X-Coord':
[0.0, 0.0], 'Znucl': 26.0, 'MULT': 2}
[ BerryPI ] ERROR: in automation of 1st_BerryPI_SG186_z0_PBE
[ BerryPI ] ERROR: Missing data in atom
[ BerryPI ] ERROR: missing tags: ('Element Name', 'Element Number')

Completed using BerryPI version: +++Version Mar 15, 2021
…

The program stops at this point:
Error in: {'Z-Coord': [0.0, 0.5], 'Y-Coord': [0.0, 0.0], 'X-Coord':
[0.0, 0.0], 'Znucl': 26.0, 'MULT': 2}

The same error was observed in the case of calculation GGA+U or in the
case of small variations of coordinate (0, 0, z) for ion Fe1.

Can somebody help me or explain the reason of such mistake?

The description for similar error was not found in the WIEN2k mailing
list. I am not sure about the reason of error, but I can suppose, that
the reason is linked with follow:   gamma= 120 degree in
1st_BerryPI_SG186_z0_PBE.struct, whereas in
1st_BerryPI_SG186_z0_PBE.outputberry gamma is 20 degree…

Thanks in advance for Your help,
Vitalii Turchenko