Problem with wannier spreads

[appalakondaiah]

Dear Samuel and Carla,

Thank you for the help in nscf.in file. Now I am facing the problem Wannier spreads. See the below for the reference. On the other hand, If I am using guiding_centres = TRUE in .win file I found 10-3 Convergence. How can I solve this in epw input !!! Below I am attaching my epw.in

Final State
WF centre and spread 1 ( -0.374968, 0.756406, -1.157746 ) 113.54692639
WF centre and spread 2 ( -0.475653, 0.086713, 0.990716 ) 107.38814105
WF centre and spread 3 ( 0.071831, -0.508931, -0.241204 ) 113.55715999
WF centre and spread 4 ( 0.179110, 0.564240, -1.278107 ) 113.99673662
WF centre and spread 5 ( -0.922056, -0.386755, -0.320907 ) 116.10952922
WF centre and spread 6 ( 0.748408, 2.486341, -7.888025 ) 79.73918437
WF centre and spread 7 ( 0.447206, 0.545389, -8.515494 ) 71.10852022
WF centre and spread 8 ( 2.065255, 1.028336,-11.713453 ) 5.72058108
WF centre and spread 9 ( 2.026548, 1.381210, 3.811993 ) 123.14589834
WF centre and spread 10 ( 0.241264, 2.410525, 10.879547 ) 21.89531700
WF centre and spread 11 ( 2.040362, 1.863725, 10.928668 ) 22.14314823
Sum of centres and spreads ( 6.047307, 10.227198, -4.504009 ) 888.35114252

Spreads (Ang^2) Omega I = 14.656869261
================ Omega D = 804.883549125
Omega OD = 68.810724138
Final Spread (Ang^2) Omega Total = 888.351142524


epw.in
--
&inputepw
prefix = 'NbSe2',
amass(1) = 92.91,
amass(2) = 78.96,
outdir = './'

elph = .true.
kmaps = .false.
epbwrite = .true.
epbread = .false.
system_2d = .true.
epwwrite = .true.
epwread = .false.

nbndsub = 11


wannierize = .true.
num_iter = 300
!dis_win_max = 5
proj(1) = 'Nb:d'
proj(2) = 'Se:p'

wdata(1) = 'bands_plot = .true.'
wdata(2) = 'begin kpoint_path'
wdata(3) = 'G 0.00 0.00 0.00 M 0.50 0.00 0.00'
wdata(4) = 'M 0.50 0.00 0.00 K 0.33 0.33 0.00'
wdata(5) = 'K 0.33 0.33 0.00 G 0.00 0.00 0.00'
wdata(6) = 'end kpoint_path'
wdata(7) = 'bands_plot_format = gnuplot'

iverbosity = 0

elinterp = .true.
phinterp = .true.

tshuffle2 = .true.
tphases = .false.

elecselfen = .false.
phonselfen = .true.

parallel_k = .true.
parallel_q = .false.

fsthick = 6 ! eV
eptemp = 0.075 ! K
degaussw = 0.025 ! eV

a2f = .true.

dvscf_dir = '../phonons/save'

nkf1 = 16
nkf2 = 16
nkf3 = 1

nqf1 = 16
nqf2 = 16
nqf3 = 1

nk1 = 8
nk2 = 8
nk3 = 1

nq1 = 8
nq2 = 8
nq3 = 1

/
10 cartesian
0.000000000 0.000000000 0.000000000 1.000000000
0.073128834 0.126662856 0.000000000 1.000000000
0.146257668 0.253325712 0.000000000 1.000000000
0.219386502 0.379988568 0.000000000 1.000000000
-0.292515336 -0.506651424 0.000000000 1.000000000
0.219386502 0.126662856 0.000000000 1.000000000
0.292515336 0.253325712 0.000000000 1.000000000
0.365644170 0.379988568 0.000000000 1.000000000
0.438773004 0.253325712 0.000000000 1.000000000
0.511901838 0.379988568 0.000000000 1.000000000

[sponce]

Dear appalakondaiah,

You can add any wannier90 command into EPW input file with:

Code: Select all

wdata(8) = 'guiding_centres = TRUE '

for example.


Best,

Samuel

[appalakondaiah]

Dear Prof. Samuel,

Thank you for the support. Now the error is changed and aborted the programme after the below step (Here I am just using Pb and MgB2 examples as a reference input)

Code: Select all

  Writing epmatq on .epb files


              band disentanglement is used:  nbndsub =   11

     Writing Hamiltonian, Dynamical matrix and EP vertex in Wann rep to file


     Reading Hamiltonian, Dynamical matrix and EP vertex in Wann rep from file


     Finished reading Wann rep data from file

     Using uniform q-mesh:   16  16   1
     Size of q point mesh for interpolation:        256
     Using uniform k-mesh:   16  16   1
     Size of k point mesh for interpolation:        512
     Max number of k points per pool:               22

     Fermi energy coarse grid =  -1.775686 eV

     ===================================================================

     Fermi energy is read from the input file: Ef =  -1.775700 eV

     ===================================================================

              ibndmin =     3  ebndmin =    -0.774
              ibndmax =    11  ebndmax =     0.215


     Number of ep-matrix elements per pool :         8019 ~=   62.65 Kb (@ 8 bytes/ DP)

     ===================================================================
     Solve anisotropic Eliashberg equations
     ===================================================================


===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 29421 RUNNING AT saint
=   EXIT CODE: 24
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================

[sponce]

Dear appalakondaiah,

Are you sure that you use the Pb example?
Those example should be 3D with 16x16x16 and not 16x16x1 grids.

Can you try to run the EPW tests?
They are located in EPW/tests/Inputs/

One test should be Pb.

The test should work. If its not the case can you report the error you have?

Best,

Samuel

[appalakondaiah]

Dear Prof. Samuel

Yes ..i followed Pb as a reference and prepare the same example for monolayer 2H-NbSe2 (as an input). Unfortunately, it was not running after below step..

Number of ep-matrix elements per pool : 8019 ~= 62.65 Kb (@ 8 bytes/ DP)

===================================================================
Solve anisotropic Eliashberg equations
===================================================================


with regards
S. Applakondaiah

[sponce]

Dear S. Applakondaiah,

Ok so MgB2 and Pb example both works fine for you? (those are both 3D examples with grids like 6x6x6 etc not like 6x6x1 as you show above).

Although theoretically 2D materials should work with EPW, I personally have never ran one (there are currently no examples/tests on 2D materials).

Could you provide me with the error message that you got ?

Did the calculations just hang?

PS: This should not be the error message:

Code: Select all

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 29421 RUNNING AT saint
=   EXIT CODE: 24
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================

You should have another message depending on what submission too you used (PBS should give you a .o and .e files for example).

Best,

Samuel

[sridhu88]

Dear S. Applakondaiah,

There was a small bug in memory allocation which can throw out-of-bounds problems for 2D materials. This was reported earlier in this forum: viewtopic.php?f=6&t=44

I'm not sure if your problem is related to this but in any case, make sure to correct that bug if you're working with 2D materials. I guess it's corrected in the trunk version.

Best,
Sridhar

[appalakondaiah]

Dear Samuel,
Unfortunately, I am using normal shell script to run my programme in workstation(not in PBS as like in HPCC). Here I puzzled with the sudden stopping of my job !!! The error is shown in below in the terminal and epw.out

forrtl: info (58): format syntax error at or near *(3x,E15.5))
forrtl: severe (62): syntax error in format, unit 60, file /../Ultrasoft/epw/lambda.phself
Image PC Routine Line Source
epw.x 0000000000C733CD Unknown Unknown Unknown
epw.x 0000000000C71ED5 Unknown Unknown Unknown
epw.x 0000000000C0A930 Unknown Unknown Unknown
epw.x 0000000000BA127F Unknown Unknown Unknown
epw.x 0000000000BA0AB2 Unknown Unknown Unknown
epw.x 0000000000BF2B9F Unknown Unknown Unknown
epw.x 000000000044E90C ephwann_shuffle_ 878 ephwann_shuffle.f90
epw.x 0000000000445B6A elphon_shuffle_wr 685 elphon_shuffle_wrap.f90
epw.x 000000000040B70C MAIN__ 139 epw.f90
epw.x 000000000040AB8C Unknown Unknown Unknown
libc.so.6 00007FF7538A8EC5 Unknown Unknown Unknown
epw.x 000000000040AA89 Unknown Unknown Unknown

Number of (k,k+q) pairs on the Fermi surface: 256 out of 256


ismear = 1 iq = 256 coord.: 0.93750 0.93750 0.00000 wt: 0.00391
-------------------------------------------------------------------
lambda( 1 )= 38.870283 gamma= 0.485154 meV omega= 1.5850 meV
lambda( 2 )= 117.176240 gamma= 9.335953 meV omega= 4.0046 meV
lambda( 3 )= 28.829672 gamma= 5.915712 meV omega= 6.4267 meV
lambda( 4 )= 0.155717 gamma= 0.221675 meV omega= 16.9275 meV
lambda( 5 )= 0.140280 gamma= 0.212280 meV omega= 17.4526 meV
lambda( 6 )= 1.180101 gamma= 3.730373 meV omega= 25.2243 meV
lambda( 7 )= 3.985938 gamma= 15.510125 meV omega= 27.9862 meV
lambda( 8 )= 1.141436 gamma= 5.054275 meV omega= 29.8542 meV
lambda( 9 )= 0.119184 gamma= 0.775632 meV omega= 36.1928 meV
lambda( tot )= 191.598850
-------------------------------------------------------------------


Number of (k,k+q) pairs on the Fermi surface: 256 out of 256


===================================================================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= PID 18136 RUNNING AT saint
= EXIT CODE: 62
= CLEANING UP REMAINING PROCESSES
= YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================



PS: Thank you Sridhar for the response. Indeed I am running trail case with 16X16X1 k set and 11 Wannier bands. So, memory shouldn't be the problem. If possible I will contact you in person for discuss more details


Sincerity
S. Appalakondaiah

[sponce]

Dear S. Appalakondaiah,

Ok, I think that I found the nasty bug.

This bug only happen because you are using the PGI compiler. If you use GCC or intel, I'm quite convince it will work.

The way to solve the bug is as follow:

In EPW/src/ modify all the occurences of

Code: Select all

myfmt = *(3x,E15.5))

by

Code: Select all

 myfmt = "(1000(3x,E15.5))"

The PGI compiler does not recognise the (quite) standard * format.

Best,

Samuel