0: ERROR: kgrid out of bounds in wfn

Submitted by farzanehsh on Mon, 02/15/2016 - 21:34

Forums 

Bug reports

Hello Developer,

I'm calculating MgO structure, I have a problem in "epsilon",
I use BGW(1.0.6) and QEspresso(5.2), the epsilon inputfile is:

--------------------epsilon.inp---------------------
epsilon_cutoff 5.0
number_bands 24
band_occupation 2*1 22*0
number_partial_occup 0
cell_slab_truncation
number_qpoints 15
begin qpoints
0.001000000 0.000000000 0.000000000 1.0 1
0.000000000 0.125000000 0.000000000 1.0 0
0.000000000 0.250000000 0.000000000 1.0 0
0.000000000 0.375000000 0.000000000 1.0 0
0.000000000 0.500000000 0.000000000 1.0 0
0.125000000 0.125000000 0.000000000 1.0 0
0.125000000 0.250000000 0.000000000 1.0 0
0.125000000 0.375000000 0.000000000 1.0 0
0.125000000 0.500000000 0.000000000 1.0 0
0.250000000 0.250000000 0.000000000 1.0 0
0.250000000 0.375000000 0.000000000 1.0 0
0.250000000 0.500000000 0.000000000 1.0 0
0.375000000 0.375000000 0.000000000 1.0 0
0.375000000 0.500000000 0.000000000 1.0 0
0.500000000 0.500000000 0.000000000 1.0 0
end

------------------------------------------------------
after running, there is an error message at "epsilon.out", which is " 0: ERROR: kgrid out of bounds in wfn"

Appreciated for your help,

Thanks,
..Farzaneh

Submitted by babarker on Mon, 02/22/2016 - 20:43

Hello Farzaneh,

I do not believe that this is a bug. The error message "kgrid out of bounds in wfn" occurs in Common/wfn_rho_vxc_io_inc.f90 and gets written out when the code finds that the Monkhorst-Pack grid is somehow less than 1x1x1.

Since, in private email, you mention only doing an SCF calculation and not an NSCF calculation to generate a WFN or WFNq file, I recommend doing an NSCF calculation (and probably a wavefunction conversion step with pw2bgw) to get WFN and WFNq files, if you did not already.

If you did an NSCF step, make sure your k-grid is MxNxP (depending on the system, you may want to make M not equal to N, etc.). If you use espresso, the "kgrid.x" utility is very useful. Also make sure that your k-grid is generated without time-reversal symmetry allowed.

Best,
Brad

Submitted by farzanehsh on Mon, 02/29/2016 - 20:55

hello Brad,
Thanks dor your comments, My problem has been solved,

Thanks

Submitted by farzanehsh on Sat, 02/27/2016 - 07:11

Hello Developers,

I wanna plot exciton WFN with use of PLotxct for silicon example in BGW,(BerkeleyGW-1.0.6/examples/DFT/silicon/...)
I got Exciton eigenvectors, WFN_fi and WFNq_fi as input files to use in Plotxct,
The Plotxct.inp was:
-------------------------------------------PLOTXCT.INP------------------------------
plot_state 20

q_shift 0.047000 0.03700 0.032000

supercell_size 10 10 10

hole_position 10.00000 10.00000 0.00000

begin lattice_vectors
0.000000 0.500000 0.500000
0.500000 0.000000 0.500000
0.500000 0.500000 0.000000
end

the outputfile :
----------------------------------------OUT.PLO---------------------------------------
Running MPI version (parallel)

Running on 24 processor(s)

Lattice vectors:
0.00000 0.50000 0.50000
0.50000 0.00000 0.50000
0.50000 0.50000 0.00000
Reading state # 20
Excitation energy = 3.33839063804600 eV
Reducing k-points to 1000
Wfn norm (restricted sum): 1.00000000000000
Wfn norm (full): 1.00000000000000
*** VERBOSE: input: calling distrib time = 08:32:50.444

Number of kpoints treated by each PE:
42 42 42 42 42 42 42 42
42 42 42 42 42 42 42 42
41 41 41 41 41 41 41 41
*** VERBOSE: input: reordering gvecs time = 08:32:50.444
*** VERBOSE: input: reading WFN_fi time = 08:32:50.446

crystal wavefunctions read from tape WFN_fi

nrk= 1000
0.0470 0.0370 0.0310
0.0470 0.0370 0.1310
0.0470 0.0370 0.2310
0.0470 0.0370 0.3310
0.0470 0.0370 0.4310
0.0470 0.0370 0.5310
0.0470 0.0370 0.6310
0.0470 0.0370 0.7310
0.0470 0.0370 0.8310
0.0470 0.0370 0.9310
0.0470 0.1370 0.0310
0.0470 0.1370 0.1310
..................
....................
fine grid nfk= 1000 ksz= 0.01659
# valence bands = 4
# cond. bands = 6
Coordinates of hole = 10.0000000000000 10.0000000000000
0.000000000000000E+000
lattice box = 10 10 10
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 2 in communicator MPI_COMM_WORLD
with errorcode -1.

after running, this error messsage appeares:

"ERROR: volume inconsistent with lattice vectors"
**************************************************
I checked lattice vectors, supercell size,.... , all are correct,

what should I do to resolve this problem???

Thanks in advance,
..F

jornada's picture

Submitted by jornada on Mon, 02/29/2016 - 09:05

You can try to run plotxct provided in BerkeleyGW 1.1-beta2. The file format for the bse calculation didn't change (as long as you don't compile with HDF5 support), and the new version doesn't ask you to input the lattice vectors any longer. Let us know if that works.

jornada's picture

Submitted by jornada on Wed, 03/02/2016 - 15:27

Right, but what I'm suggesting is for you to compile BGW 1.1-beta2...

You could of course figure out how to write the lattice vectors. I never used that very old PlotXct, but I think you problem is that you need to specify the lattice vector in either angstroms or in bohrs. But most definitely, your lattice vectors should *not* be in units of "alat" like in QE.

But again, just compile & run 1.1-beta2 and save yourself of all this trouble ;)

Cheers,
Felipe

Submitted by babarker on Thu, 03/03/2016 - 11:28

Hello Farzaneh,

How did you fix the original problem? Was it a matter of getting a good kgrid for the NSCF calculations?

Best,
Brad