problem by ypp -e a

[xixi]

Hi ,
when I do a pwscf calculation, I use a wrong input file, which includes finite electric field ( I am not sure whether yambo can handle this, just try),
than I continue to do the next,
when I use ypp -e a, I find the file o.exc_weights_at_1 is very strange

===================================================
#
# Band_V Band_C K ibz Symm. Weight Energy
#
#
# YPP@mse.ntu.edu.sg x 001 CPUs * 03/11/2010 18:26 [start]
# 03/11/2010 18:26 [end]
==================================================================
the state 1 has the biggest strength

there should be some band index below the "Band_V Band_C' row to give some information about the states "1 - 1" ,
but there isn't.
but for some other states, which isn't of the big strength, it's fine.

does it mean the exciton may be ionized by the E-field for the state "1-1" and ypp can't index the related bands ?

Thanks!

Xi Zhu

School of Materials Science And Engineering,
Nanyang Technological University,
Blk N4.1, Nanyang Avenue,
Singapore 639798

[Daniele Varsano]

Dear Xi Zhu,

which includes finite electric field ( I am not sure whether yambo can handle this, just try),

Yes, Yambo can handle it, but of course this influence all the calculation, so if you don't want it, you
should not include it.

Anyway,
the problem you are facing, should be not related to the finite electric field.
Remember that you should uncomment the flag:
WRbsWF # [BSS] Write to disk excitonic the FWs
and use the diagonalization method as a solver.
I cannot say what's your problem now, are you sure that's' index 1 is your most intense
excitations. Try to look at the excitation order by typing
> ypp -e s
and look at the sorted output file, and eventually post them here together with your input file.

In the o.exc_weights_at_1 you have reported the
# Electron-Hole pairs that contributes to Excitonic State 1 more than 5%
and may be none of the electron-hole pairs you have satisfy this statement,
you can lower this threshold with a small change in the code.

Hope it helps,

Daniele

[bob]

Hi people,

I do have a very similar problem.
After having solved the BSE using

yambo -o b -y d

I sort the excitons "ypp -e s" and get the biggest strength at index 23. If I want to plot the wave function with "ypp -e w" or the amplitude with "ypp -e a", I get a segmentation fault at

Code: Select all

<---> [01] Y(ambo) P(ost)/(re) P(rocessor)
 <---> [02] Core DB
 <---> :: Electrons             : 196.0000
 <---> :: Temperature       [ev]: 0.000000
 <---> :: Lattice factors [a.u.]: 69.91987  47.24315  51.02261
 <---> :: K points              : 1
 <---> :: Bands                 :  500
 <---> :: Symmmetries           : 2
 <---> :: RL vectors            :  90067
 <---> [03] K-point grid
 <---> :: Q-points   (IBZ): 1
 <---> :: X K-points (IBZ): 1
 <---> [04] CORE Variables Setup
 <---> [04.01] Unit cells
 <---> [04.02] Symmetries
 <---> [04.03] RL shells
 <---> [04.04] K-grid lattice
 <---> [04.05] Energies [ev] & Occupations
 <---> [05] Excitonic Properties
 <---> :: Sorting energies
 <---> [05.01] Excitonic Amplitude
 <---> Processing 1 states
 <---> [05.02] Reporting amplitude and weightsSegmentation fault

I do have "WRbsWF" in the yambo inuput file.

Any clue what that problem may be related to?

[Daniele Varsano]

Dear Bjoern,
no idea on what's happening. Anyway, is some output (o.exc_weight, o.exc_amplitude) produced
before crashing?
you can have a look at ./ypp/excitons_sort_and_report.F subroutine to localize the problem.
Otherwise we should reproduce your error in order to individuate the problem.
Let us know,

Cheers,

Daniele

[bob]

Well, the o.exc_amplitude_at_23 file only contains the header

Code: Select all

#  __  __   ________   ___ __ __    _______   ______        
# /_/\/_/\ /_______/\ /__//_//_/\ /_______/\ /_____/\       
# \ \ \ \ \\::: _  \ \\::\| \| \ \\::: _  \ \\:::_ \ \      
#  \:\_\ \ \\::(_)  \ \\:.      \ \\::(_)  \/_\:\ \ \ \     
#   \::::_\/ \:: __  \ \\:.\-/\  \ \\::  _  \ \\:\ \ \ \    
#     \::\ \  \:.\ \  \ \\. \  \  \ \\::(_)  \ \\:\_\ \ \   
#      \__\/   \__\/\__\/ \__\/ \__\/ \_______\/ \_____\/   
#                                                           
#             GPL Version 3.2.3 Revision 676
#                http://www.yambo-code.org
#
# BSK|Identifier              :1161
#    |Dimension               :420
#    |Bands                   :78 - 118
#    |Exchange           [res]: yes
#    |Correlation        [res]: yes
#    |Kernel`s coupling       : yes
#    |Exchange           [cpl]: yes
#    |W interaction is bare   : no
#    |Correlation        [cpl]: yes
#    |ALDA kernel in R-space  : no
#    |RL vectors    [exchange]:9999
#    |RL vectors [correlation]:4007
#    |E/h energy range    [ev]:-1.000000 - -1.000000
#    |Coupling range     [o/o]: 100.0000 -  100.0000
#  W |Interaction is diagonal : no
#    |Matrix size             :4007
#    |Bands                   :1 - 500
#    |e/h energy range    [ev]:-1.000000 - -1.000000
#    |Poles              [o/o]: 100.0000
#    |Rl vectors in the sum   :9999
#    |[r,Vnl] included        : no
#    |Field direction         :0.1000E-4  0.000000  0.000000
#    |Coulomb Cutoff        :none
#    |xc-Kernel               :none
# RIM|RL components      [col]:0
#    |Random points      [col]:0
#
# Electron-Hole pairs that contributes to Excitonic State 23 more than 5%
#

I'm a bit puzzled why it says

Code: Select all

#    |E/h energy range    [ev]:-1.000000 - -1.000000
#    |Coupling range     [o/o]: 100.0000 -  100.0000

Looking at excitons_sort_and_report.F, it seems like the error occurs early in this loop:

Code: Select all

   K_weight=0.
   do neh = 1,BS_H_dim
       ikbz  = BS_eh_table(S_indx(neh),1)
       ikibz = Xk%sstar(ikbz,1)
     K_weight(ikibz)=K_weight(ikibz)+A_weight(neh)
   enddo

In the first pass, ikbz is 844639183 (?) and then the segmentation fault is when ikibz = Xk%sstar(ikbz,1).

[andrea marini]

Dear Bjoern,

to trace back the problem and to be sure that it is really a bug I need to reproduce the error. Any simple test I have done did not show any seg fault. Is there any way we can reproduce EXACTLY you run ? Can you provide us the NETCDF databases and/or the ground state input files, pseudo and so on ? If the databases are too large we can use rapidshare to exchange them.

Let us know.

Andrea

[Daniele Varsano]

Dear Bjoern,

Code: Select all

#    |E/h energy range    [ev]:-1.000000 - -1.000000
#    |Coupling range     [o/o]: 100.0000 -  100.0000

do not worry about that!. You can choose the band to include in your BSE matrix by giving
e-h energies, here is negative, so yambo read the number of bands (78-118) in your case.

Here is surely the problem:
ikbz is 844639183 (?)

you are working at gamma point right?, so it should be 1.
It looks that the table of the BS matrix are wrong.

Could you post the BSE report, as well as the o.exc_E/I outputs and the ypp.in?
Does it happens also for other exciton indexes?

Let's have a look to the reports if we can find some hints on what's going wrong, otherwise
we should try to reproduce you error.

Cheers,

Daniele

[bob]

Dear Daniele, dear Andrea,

Thanks for your replies!I tried the same calculations for a small C2H4 molecule and no error occurred. I attached the files you asked for resulting in the error, and here is ypp.in

Code: Select all

#   __   __     _        __  __       ____      U  ___  u
#   \ \ / / U  /"\  U u |" \/ "| u U | __") u    \/"_ \/
#    \ V /   \/ _ \/   \| |\/| |/   \|  _ \/     | | | |
#   U_|"|_u  / ___ \    | |  | |     | |_) | .-,_| |_| |
#     |_|   /_/   \_\   |_|  |_|     |____/   \_)-\___/
# .-,//|(_   \\    >>  <<,-,,-.     _|| \\_        \\
#  \_) (__) (__)  (__)  (./  \.)   (__) (__)      (__)
#
#             GPL Version 3.2.3 Revision 676
#                http://www.yambo-code.org
#
wavefunction                 # [R] Wavefunction
excitons                     # [R] Excitons
Format= "x"                  # Output format [(c)ube/(g)nuplot/(x)crysden]
Direction= "123"               # [rlu] [1/2/3] for 1d or [12/13/23] for 2d [123] for 3D
FFTGvecs=  9999        RL    # [FFT] Plane-waves
States= "23"                 # Index of the BS state(s)
Degen_Step=   0.0100   eV    # Maximum energy separation of two degenerate states
% Cells
 1 | 1 | 1 |                             # Number of cell repetitions (even or 1)
%
% Hole
 0.000    | 0.000    | 0.000    |        # [cc] Hole position in unit cell
%

Can you see anything from those files?

Cheers,
Bjoern

[Daniele Varsano]

Dear Bjoern,

the input you sent me is for the plot, and not for the weight, anyway it doesn't matter.
The report is from a previous calculations, in the report all the databases are read, anyway
everything looks went ok!!

Anyway, please try to put in the ypp.in:
States= "23 - 23"

instead of just 23.
It could be the problem!!!

Let me know,

Daniele

[bob]

Dear Bjoern,

the input you sent me is for the plot, and not for the weight, anyway it doesn't matter.
The report is from a previous calculations, in the report all the databases are read, anyway
everything looks went ok!!

Did you mean the input for ypp -e s?

Anyway, please try to put in the ypp.in:
States= "23 - 23"

instead of just 23.
It could be the problem!!!

Sadly, it's not. I had 23 - 23 before and changed to 23 just for trying that out.