NetCDF: NC_UNLIMITED

Various technical topics such as parallelism and efficiency, netCDF problems, the Yambo code structure itself, are posted here.

Moderators: Daniele Varsano, andrea.ferretti, andrea marini, Conor Hogan, myrta gruning

NetCDF: NC_UNLIMITED

Postby deliciousnai » Mon Jul 03, 2017 11:19 am

Dear Developers,

I want to study optical spectrum with GW-BSE correction in pure GaN and AlGaN with ratio Al to Ga 1:7, 1:3 and 1:1. I don't have any problem with pure GaN and 1:1 AlGaN (maybe there is some physical problem which I can consult with my professor), however, I got some problem when I calculate the 1:7 AlGaN.

[ERROR] STOP signal received while in :[05] Dynamic Dielectric Matrix (PPA)
[ERROR][NetCDF] NetCDF: NC_UNLIMITED in the wrong index


I use Yambo 4.1.1 with yambo-py help to generate input (because manual converge test is such a pain), I wonder if the problem is insufficient memory?

Since my zipped QE input files (scf and nscf) and GW-converge input files exceed allowed size, please download the tar file here:
https://drive.google.com/open?id=0B_FpAew-Eh53NUkyemxsaHl4Wmc

Another question for 1:1 AlGaN, will it be a problem if I use same input with GaN? I suppose since the QE input almost same, the yambo-py will generate same yambo input (the number of k points and number atom exactly same, since Ga and Al have same electron valence hence they have a same number of bands too), but I keep wondering if its alright or not.

Regards,

Naima.
Last edited by deliciousnai on Tue Jul 04, 2017 5:34 am, edited 1 time in total.
Naima Amaliah
Physics Department Undergraduate Student
University of Indonesia
deliciousnai
 
Posts: 10
Joined: Fri May 26, 2017 11:47 am
Location: Indonesia

Re: NetCDF: NC_UNLIMITED

Postby Daniele Varsano » Mon Jul 03, 2017 11:37 am

Dear Naima,
can you post just the input/report/output of the run where you encounter the problem?

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/
User avatar
Daniele Varsano
 
Posts: 1974
Joined: Tue Mar 17, 2009 2:23 pm

Re: NetCDF: NC_UNLIMITED

Postby deliciousnai » Tue Jul 04, 2017 5:34 am

Dear Daniele,

for nscf input:

Code: Select all
&control
         calculation = 'nscf'
              prefix = 'gan'
          pseudo_dir = './pseudos'
         outdir = './outdir'
           verbosity = 'high'
          wf_collect = .true.
/&end
&system
                   A = 6.378
                   B = 6.378
                   C = 5.185
               cosAB = -0.5
               cosBC = 0
               cosAC = 0
             ecutwfc = 60
    force_symmorphic = .true.
         occupations = 'fixed'
               ibrav = 4
                 nat = 16
                nbnd = 60
                ntyp = 3
/&end
&electrons
      diago_full_acc = .true.
            conv_thr = 1e-08
/&end
ATOMIC_SPECIES
   Al   26.98154  Al.pbe-mt_fhi.UPF
  Ga   69.723    Ga.pbe-mt_fhi.UPF
   N 14.00674     N.pbe-mt_fhi.UPF
ATOMIC_POSITIONS { crystal }
Ga       0.000924955   0.001801040  -0.003730560
Al       0.166683664   0.333316336   0.502241014
Ga       0.500016368  -0.000016368  -0.004029008
Ga       0.667018042   0.333984518   0.494987756
Ga       0.000925105   0.499074895  -0.003731056
Ga       0.166015482   0.832981958   0.494987756
Ga       0.498198960   0.499075045  -0.003730560
Ga       0.667018479   0.832981521   0.494987166
N        0.004535075   0.009021145   0.381205361
N        0.166684222   0.333315778   0.868124848
N        0.500016156  -0.000016156   0.378497621
N        0.667110531   0.334170673   0.879260063
N        0.004534069   0.495465931   0.381204811
N        0.165829327   0.832889469   0.879260063
N        0.490978855   0.495464925   0.381205361
N        0.667110710   0.832889290   0.879259363
K_POINTS { automatic }
 12 12  1  0  0  0


here one of the input for my converge test (BndsRnXp= 1 | 10)

Code: Select all
em1d
gw0
HF_and_locXC
ppa
% XfnQP_Wv
 0.0 | 0.0 | 0.0 |   
%
BoseCut = 0.100000
GfnQP_Wc_E = 0.000000 eV
% XfnQP_Wc
 0.0 | 0.0 | 0.0 |   
%
DBsFRAGpm =     'none'
% QpntsRXp
 1 | 74 |   
%
BoseTemp = -1.000000 eV
ElecTemp = 0.000000 eV
NGsBlkXp = 0 Ry
% LongDrXp
 1.0 | 0.0 | 0.0 |   
%
CGrdSpXp = 100.000000
NLogCPUs = 0.000000
FFTGvecs = 10921.000000 RL
StdoHash = 40.000000
XTermKind =     'none'
GDamping = 0.100000 eV
DBsIOoff =     'none'
GfnQP_Wv_E = 0.000000 eV
% GbndRnge
 1 | 10 |   
%
XfnQP_Wc_E = 0.000000 eV
% GfnQP_Wv
 0.0 | 0.0 | 0.0 |   
%
% EhEngyXp
 -1.000000 | -1.000000 |  eV
%
dScStep = 0.100000 eV
% QPkrange
 74 | 74 | 4 | 5 |   
%
Nelectro = 64.000000
% BndsRnXp
 1 | 10 |   
%
DysSolver =        'n'
XfnQPdb =     'none'
% GfnQP_Wc
 0.0 | 0.0 | 0.0 |   
%
GfnQPdb =     'none'
EXXRLvcs = 10 Ry
kind =       'BG'
XfnQP_Z = (1.000000,0.000000)
XTermEn = 40.817080 eV
XfnQP_Wv_E = 0.000000 eV
% XfnQP_E
 0.0 | 1.0 | 1.0 |   
%
GfnQP_Z = (1.000000,0.000000)
GTermEn = 40.817080 eV
OccTresh = 0.000010
% GfnQP_E
 0.0 | 1.0 | 1.0 |   
%
GTermKind =     'none'
XfnQP_N = 1.000000
GfnQP_N = 1.000000
PPAPntXp = 27.211380 eV
X_all_q_nCPU_LinAlg_INV = 1.000000


the log file and report file for this input is in the attachment

I also did converge test for EXXRLvcs, GbndRnge and NGsBlkXp (just like the tutorial suggest) and all of them gave almost same result at last row:

[ERROR] STOP signal received while in :[05] Dynamic Dielectric Matrix (PPA)
[ERROR][NetCDF] NetCDF: NC_UNLIMITED in the wrong index


Regards,

Naima.
You do not have the required permissions to view the files attached to this post.
Naima Amaliah
Physics Department Undergraduate Student
University of Indonesia
deliciousnai
 
Posts: 10
Joined: Fri May 26, 2017 11:47 am
Location: Indonesia

Re: NetCDF: NC_UNLIMITED

Postby Daniele Varsano » Tue Jul 04, 2017 9:19 am

Dear Naima,
I can see a couple of inconsistencies in your input file:

1)
Code: Select all
 kind =       'BG'

There is no variable named just kind in Yambo, probably here you wanted to use the XTermKind, this activates the terminator technique in the response function in order to accelerate in any case I suggest you to not use it if it is not needed and you do not know exactly what you are doing.
Moreover:
2)
Code: Select all
NGsBlkXp = 0 Ry
this does not make sense. Set a valid number either in Ry or in RL.

I suggest you to remove old databases ndb.*
and generate your input file from scratch by using yambo command lines:
yambo -x -p p -g n
edit your inout variables and run yambo, post it again your input/report/output file if the problem persists.

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/
User avatar
Daniele Varsano
 
Posts: 1974
Joined: Tue Mar 17, 2009 2:23 pm

Re: NetCDF: NC_UNLIMITED

Postby deliciousnai » Tue Jul 04, 2017 12:40 pm

Dear Daniele,

I generate input using yambo -x -g n -p p and put NGsBlkXp = 10 Ry
and the same problem still exist...

I attached the new input, log file and output

Regards,
Naima.
You do not have the required permissions to view the files attached to this post.
Naima Amaliah
Physics Department Undergraduate Student
University of Indonesia
deliciousnai
 
Posts: 10
Joined: Fri May 26, 2017 11:47 am
Location: Indonesia

Re: NetCDF: NC_UNLIMITED

Postby Daniele Varsano » Tue Jul 04, 2017 2:00 pm

Dear Naima,

you have 32 occupied states:
Code: Select all
States summary         : Full        Metallic    Empty
                           0001-0032               0033-0060


The response function is built by considering electron-hole transitions, or if you want considering pairs of occupied and empty states.
So at least 1 unoccupied state need to be cosidered and this is not the case wit your input file:
Code: Select all
% BndsRnXp
   1 |  10 |                 # [Xp] Polarization function bands
%


So raise the number of bands including some unoccupied state. Next the number of bands it is a parameter to be converged.
Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/
User avatar
Daniele Varsano
 
Posts: 1974
Joined: Tue Mar 17, 2009 2:23 pm

Re: NetCDF: NC_UNLIMITED

Postby deliciousnai » Wed Jul 05, 2017 4:39 am

Dear Daniele,

Thank you, it works now :D

Regards,
Naima.
Naima Amaliah
Physics Department Undergraduate Student
University of Indonesia
deliciousnai
 
Posts: 10
Joined: Fri May 26, 2017 11:47 am
Location: Indonesia


Return to Technical Issues

Who is online

Users browsing this forum: No registered users and 1 guest

cron