how to solve the problem during the GW calculation?

[artingchen]

Hi, veryone,
I am a new YAMBO user. I recently used the YAMBO to calculate the bandstructure fo In203, there was a error as follows:

<---> [01] Files & I/O Directories
<---> [02] CORE Variables Setup
<---> [02.01] Unit cells
<---> [02.02] Symmetries
<---> [02.03] RL shells
<---> [02.04] K-grid lattice
<---> [02.05] Energies [ev] & Occupations
<01s> [WARNING]Non zero Temp. is still unsupported in this version
<01s> [03] Transferred momenta grid
<01s> [04] Bare local and non-local Exchange-Correlation
<01s> [M 3.805 Gb] Alloc WF ( 3.793)
<03s> [FFT-HF/Rho] Mesh size: 40 40 40
<03s> [WF-HF/Rho loader] Wfs (re)loading | | [000%] --(E) -- <08s> [WF-HF/Rho loader] Wfs (re)loading |## | [012%] 05s(E) 3 <13s> [WF-HF/Rho loader] Wfs (re)loading |##### | [026%] 10s(E) 3 <18s> [WF-HF/Rho loader] Wfs (re)loading |####### | [039%] 15s(E) 3 <23s> [WF-HF/Rho loader] Wfs (re)loading |########## | [052%] 20s(E) 3 <28s> [WF-HF/Rho loader] Wfs (re)loading |############# | [065%] 25s(E) 3 <33s> [WF-HF/Rho loader] Wfs (re)loading |############### | [079%] 30s(E) 3 <38s> [WF-HF/Rho loader] Wfs (re)loading |################## | [092%] 35s(E) 3 <41s> [WF-HF/Rho loader] Wfs (re)loading |####################| [100%] 37s(E) 37s(X)
<44s> EXS | | [000%] --(E) --(X)[hpc:08950] *** Process received signal ***
[hpc:08950] Signal: Segmentation fault (11)
[hpc:08950] Signal code: Address not mapped (1)
[hpc:08950] Failing at address: 0x4
[hpc:08950] [ 0] /lib64/libpthread.so.0 [0x33a520e7c0]
[hpc:08950] [ 1] /home/ms/yambo-3.2.5/bin/yambo(xco_hartree_fock_+0x108b) [0x46bdcb]
[hpc:08950] [ 2] /home/ms/yambo-3.2.5/bin/yambo(xco_driver_+0x3f7) [0x469ca7]
[hpc:08950] [ 3] /home/ms/yambo-3.2.5/bin/yambo(yambo_driver_+0x470) [0x426a30]
[hpc:08950] [ 4] /home/ms/yambo-3.2.5/bin/yambo(main+0xb67) [0x424127]
[hpc:08950] [ 5] /lib64/libc.so.6(__libc_start_main+0xf4) [0x33a461d994]
[hpc:08950] [ 6] /home/ms/yambo-3.2.5/bin/yambo [0x4230c9]
[hpc:08950] *** End of error message ***
Segmentation fault

Before GW calculation, I have done pwscf calculation. and get the bandstrcuture (see the figure of bandstructure). the imput files both of pwscf and yambo are attached.
can anyone help me to solve this problem? any suggestion is appreicated ! thank you!

[claudio]

Dear Aqing Chen

please put your PWSCF and YAMBO input in attachment, so we will have a look

regards
Claudio Attaccalite

[artingchen]

Dear Aqing Chen

please put your PWSCF and YAMBO input in attachment, so we will have a look

regards
Claudio Attaccalite

Dear Claudio Attaccalite,

I am sorry for the unclear attachment files' name. I have changed them. the PWSCF and YAMBO input files have been attached, you will find it.

[claudio]

Daer Aqing Chen

The problem is in the k-point sampling you used for the PWSCF calculation.

I advice you to make the following changes in your pwscf input:

1) use calculation = 'nscf' instead of calculation = 'bands'

2) added the flag force_symmorphic=.true. to the system section because Yambo doesn't support
non-symmorphic symmetries

3) use an automatic generated k-point sampling as for instance:
K_POINTS automatic
2 2 2 0 0 0
(then you have to check convergence on k-points and bands)

with best regards
Claudio Attaccalite

[artingchen]

Daer Aqing Chen

The problem is in the k-point sampling you used for the PWSCF calculation.

I advice you to make the following changes in your pwscf input:

1) use calculation = 'nscf' instead of calculation = 'bands'

2) added the flag force_symmorphic=.true. to the system section because Yambo doesn't support
non-symmorphic symmetries

3) use an automatic generated k-point sampling as for instance:
K_POINTS automatic
2 2 2 0 0 0
(then you have to check convergence on k-points and bands)

with best regards
Claudio Attaccalite

Dear Claudio Attaccalite,

Thank you for your reply, your suggestions are very useful. But I have another question. If I use an automatic generated K-piont sampling, how can I obtain the special
K-pionts ? such as the bandstructure shown in my first post.

[claudio]

Dear Aqing Chen

Thank you for your reply, your suggestions are very useful. But I have another question. If I use an automatic generated K-piont sampling, how can I obtain the special
K-pionts ? such as the bandstructure shown in my first post.

this question is discussed in the Yambo documentation:

How to calculate GW correction in a generic k-point
http://www.yambo-code.org/doc/inputs/ypp_kk.php

best regards
Claudio

[artingchen]

Daer Aqing Chen

The problem is in the k-point sampling you used for the PWSCF calculation.

I advice you to make the following changes in your pwscf input:

1) use calculation = 'nscf' instead of calculation = 'bands'

2) added the flag force_symmorphic=.true. to the system section because Yambo doesn't support
non-symmorphic symmetries

3) use an automatic generated k-point sampling as for instance:
K_POINTS automatic
2 2 2 0 0 0
(then you have to check convergence on k-points and bands)

with best regards
Claudio Attaccalite

Dear Claudio Attaccalite,

Thank you for your reply, your suggestions are very useful. But I have another question. If I use an automatic generated K-piont sampling, how can I obtain the special
K-pionts ? such as the bandstructure shown in my first post.

Dear Claudio,
your are always kindly to reply to me. your suggestions are very useful. I followed the method you told me and calculated my case again. It is fine until the step 6 where another problem occured. that's

[06] Dyson equation: Newton solver
==================================

[Newton] SC iterations :0


[06.01] G0W0 : the P(lasmon) P(ole) A(pproximation)
===================================================

[GW/PPA] Bands range : 1 200
[GW/PPA] G damping [ev]: 0.10000

QP @ K 001 - 014 : b 001 - 200

[FFT-SC] Mesh size: 36 36 36

[RD./SAVE//ndb.pp]------------------------------------------
Brillouin Zone Q/K grids (IBZ/BZ): 14 27 14 27
RL vectors (WF): 20287
Coulomb cutoff potential :none
Electronic Temperature [K]: 4.641803
Bosonic Temperature [K]: 4.641803
PPA diel. fun. energies :perdew & Zunger (xc)
wavefunctions :perdew & Zunger (xc)
Global Gauge :length
X matrix size : 1
X band range : 1 200
X e/h energy range [ev]:-1.000000 -1.000000
X Time ordering :c
X xc-Kernel :none
X Drude frequency : 0.00 0.00
X poles [o/o]: 100.0000
Rl vectors in the sum : 20287
[r,Vnl] included :no
Field direction :0.1000E-4 0.000 0.000
BZ energy Random IM :no
BZ energy RIM points :0
PPA Im energy [ev]: 27.21138
- S/N 000157 ---------------------------- v.03.02.05 r.*** -

[ERROR] STOP signal received while in :[06.01] G0W0 : the P(lasmon) P(ole) A(pproximation)

[ERROR][NetCDF] NetCDF: Unknown file format

At the first, I thought the code don't complied correctly. But I calculated the example such the sample of Si, it is ok. I don't know how to solve it. I attached the input file of pwscf and yambo. I hope you can help me again. Thank you very much!

yours

Aqing chen

[claudio]

Dear Aqing Chen

I don't know how you get that error.
I tested that code and eveythong works, I send you my input files.

In order you have to do:

1) ITO.scf.in and ITO.nscf.in for the Kohn-Sham eigenvalues, egigenvectors
2) read the WF with p2y
3) run the setup yambo -F yambo.in_setup
4) run GW with yambo -F yambo.in_gw
5) shift you grid in another point with ypp -F ypp.in
6) make another folder SHIFT, go in that folder and run again ITO.scf.in and then ITO.nscf_shift.in
7) read the new wavefunctions with p2y
8) run the setup and specify the size of the regular grid (the number of k-points of the first calculation) with yambo -F yambo.in_setup_shift
9) copy the ndb.pp file from your first calculation to the new SAVE falder and run yambo -F yambo.in_gw_shift

rember to check all the convergence parameters as k-point grid size, Response block size, number of conduction bands and so on.
Last thing in you input for the non-self-consistent calculation the parameter for the convergence of the bands in pw.x is diago_thr_init = 1.D-6
and not conv_thr that is used in the SCF cycle.

regards
Claudio Attaccalite

[artingchen]

Dear Aqing Chen

I don't know how you get that error.
I tested that code and eveythong works, I send you my input files.

In order you have to do:

1) ITO.scf.in and ITO.nscf.in for the Kohn-Sham eigenvalues, egigenvectors
2) read the WF with p2y
3) run the setup yambo -F yambo.in_setup
4) run GW with yambo -F yambo.in_gw
5) shift you grid in another point with ypp -F ypp.in
6) make another folder SHIFT, go in that folder and run again ITO.scf.in and then ITO.nscf_shift.in
7) read the new wavefunctions with p2y
8) run the setup and specify the size of the regular grid (the number of k-points of the first calculation) with yambo -F yambo.in_setup_shift
9) copy the ndb.pp file from your first calculation to the new SAVE falder and run yambo -F yambo.in_gw_shift

rember to check all the convergence parameters as k-point grid size, Response block size, number of conduction bands and so on.
Last thing in you input for the non-self-consistent calculation the parameter for the convergence of the bands in pw.x is diago_thr_init = 1.D-6
and not conv_thr that is used in the SCF cycle.

regards
Claudio Attaccalite

Dear Claudio,

Thanks for your patient reply. you are so kind. I will test it again.