K-grid is not fully q-connected @ Q**

Concerns any physical issues arising during the setup step (-i option). This includes problems with symmetries, k/q-point sets, and so on. For technical problems (running in parallel, etc), refer to the Technical forum.

Moderator: Daniele Varsano

K-grid is not fully q-connected @ Q**

Postby elena.cannuccia » Wed May 24, 2017 12:38 pm

Dear all,

I aim to calculate the elph renormalization of the energy band gap of 4H SiC.
I am following the procedure:
    1) scf
    2) nscf (for including more bands)
    3) Dvscf, reading the "scf".save
    3) Elph matrix elements, reading the "nscf".save

Whenever I am about to create the "ndb.elph*" for yambo_ph, I got stuck because of the warning mentioned in the title.
Can you help me to understand and how to solve that? Is that related to the fractional translations present in this system?

This warning comes out for every k and q grid except the 3x3x1.
I tried 10x10x3, 8x8x2 and 6x6x2, with the same "negative" result.

I am using the 4.1.3 Yambo version and QE 6.1.

Elena Cannuccia
Aix-Marseille Universiy
Dip. Fisica Univ. Roma Tor Vergata
You do not have the required permissions to view the files attached to this post.
elena.cannuccia
 
Posts: 17
Joined: Tue Oct 11, 2016 10:58 am

Re: K-grid is not fully q-connected @ Q**

Postby andrea marini » Wed May 24, 2017 1:22 pm

Ciao,

in the last revisions of the Yambo, in order to folow the rapid development of the QE code and the related e-p interface the whole procedure to get the YAMBO e-p databases has been deeply revised.

We will soon add an updated section about the proper PW procedure in the on-line tutorial

http://www.yambo-code.org/tutorials/Ele ... /index.php

Stay tuned.

In the meantime we will also try to have a look at your zip file.

Take care

Andrea
Andrea MARINI
Istituto di Struttura della Materia, CNR, (Italy)
User avatar
andrea marini
 
Posts: 447
Joined: Mon Mar 16, 2009 4:27 pm

Re: K-grid is not fully q-connected @ Q**

Postby Davide Sangalli » Wed May 24, 2017 4:59 pm

Ciao Elena,
when generating the elph matrix elements, you are using the ldist=.true. option in pwscf, right ?
In that case ph.x defines a q grid which maybe in general different from the one which yambo defines.
To avoid the error message you should let yambo read the q-points defined by ph.x when doing the initialization.

It is already coded.
Just copy the s.dbph_* databases inside the SAVE folder and do the setup with yambo_ph.
If you already did a setup, first remove the ndb.* generated by the previous setup.
When yambo reads properly the q-points from the s.dbph_*, you get a message during the log
Code: Select all
 <---> [03] Transferred momenta grid
 <---> [RL indx] Q-grid:Electron-Phonon databases



It maybe that the s.dbph_* are not found however.
In that case you should open the file "src/interface/INIT_q_points.F" and edit line 126, replacing
Code: Select all
     write (db_name,'(a,i6.6)') 's.dbph_',iabs(i_q)

with
Code: Select all
     write (db_name,'(a,i6.6)') 'SAVE/s.dbph_',iabs(i_q)

Recompile yambo_ph and then do the setup again (always remember to eventually erase the ndb.* files from previous setup).

After that ypp_ph should work without error message.

D.
Davide Sangalli, PhD
CNR-ISM, Division of Ultrafast Processes in Materials (FLASHit) and MaX Centre
http://www.ism.cnr.it/en/davide-sangalli-cv/
http://www.max-centre.eu/
User avatar
Davide Sangalli
 
Posts: 316
Joined: Tue May 29, 2012 4:49 pm
Location: Via Salaria Km 29.3, CP 10, 00016, Monterotondo Stazione, Italy

Re: K-grid is not fully q-connected @ Q**

Postby elena.cannuccia » Wed May 24, 2017 9:40 pm

Ciao Davide,

thank you for your kind reply. Unfortunately this does not cure the problem.

In my case the s.db* are correctly read and the the q-points are correctly generated without modifying the code.

The warning comes out at ypp_ph level: in the file ypp/elph/ELPH_database.F at the line 234 you have
Code: Select all
 if (io_err==-4) call warning("K-grid is not fully q-connected @ Q"//trim(intc(iq)))


some lines below (598)
Code: Select all
 if (any((/k_found==0/))) ELPH_databases_io=-4


Do you have any idea ?

Elena
elena.cannuccia
 
Posts: 17
Joined: Tue Oct 11, 2016 10:58 am

Re: K-grid is not fully q-connected @ Q**

Postby elena.cannuccia » Wed May 24, 2017 9:46 pm

In order to be more clear I will post the output of ypp_ph:

<---> [05.04] K-grid lattice
<01s> [05.05] Energies [ev] & Occupations
<01s> [06] == Electron-Phonon Interface: PW->Yambo Databases ==
<01s> PW(ELPH) databases ...[PHONON] ...found 14 Q-grid compatible
<01s> ELPH databases (WRITE) | | [000%] --(E) --(X)
<01s> [WARNING]K-grid is not fully q-connected @ Q1
<01s> [WARNING]K-grid is not fully q-connected @ Q2
<01s> [WARNING]K-grid is not fully q-connected @ Q3
<01s> [WARNING]K-grid is not fully q-connected @ Q4
<01s> [WARNING]K-grid is not fully q-connected @ Q5
<02s> [WARNING]K-grid is not fully q-connected @ Q6
<02s> [WARNING]K-grid is not fully q-connected @ Q7
<02s> [WARNING]K-grid is not fully q-connected @ Q8
<02s> [WARNING]K-grid is not fully q-connected @ Q9
<02s> [WARNING]K-grid is not fully q-connected @ Q10
<03s> [WARNING]K-grid is not fully q-connected @ Q11
<03s> [WARNING]K-grid is not fully q-connected @ Q12
<03s> [WARNING]K-grid is not fully q-connected @ Q13
<03s> ELPH databases (WRITE) |########################################| [100%] 02s(E) 02s(X)
<03s> :: Modes : 24
<03s> :: Bands range : 18
<03s> [07] Game Over


Elena
elena.cannuccia
 
Posts: 17
Joined: Tue Oct 11, 2016 10:58 am

Re: K-grid is not fully q-connected @ Q**

Postby Davide Sangalli » Thu May 25, 2017 8:05 am

Ciao Elena,
from your log I see ypp is reading the databases when running the expansion runlevel. Are you sure also yambo is reading the elph databases when doing the setup?

Try to post the logfile from the setup.

Best,
D.
Davide Sangalli, PhD
CNR-ISM, Division of Ultrafast Processes in Materials (FLASHit) and MaX Centre
http://www.ism.cnr.it/en/davide-sangalli-cv/
http://www.max-centre.eu/
User avatar
Davide Sangalli
 
Posts: 316
Joined: Tue May 29, 2012 4:49 pm
Location: Via Salaria Km 29.3, CP 10, 00016, Monterotondo Stazione, Italy

Re: K-grid is not fully q-connected @ Q**

Postby elena.cannuccia » Mon May 29, 2017 9:24 am

Ciao Davide,

as promised here the log file from setup

<01s> [01] CPU structure, Files & I/O Directories
<01s> [02] CORE Variables Setup
<01s> [02.01] Unit cells
<01s> [02.02] Symmetries
<01s> [02.03] RL shells
<01s> Shells finder |########################################| [100%] --(E) --(X)
<01s> [02.04] K-grid lattice
<01s> [02.05] Energies [ev] & Occupations
<01s> [03] Transferred momenta grid
<01s> [RL indx] Q-grid:Electron-Phonon databases
<01s> X indexes |########################################| [100%] --(E) --(X)
<01s> SE indexes |########################################| [100%] --(E) --(X)
<01s> [04] External corrections
<01s> [05] Game Over & Game summary

Thank you!
elena.cannuccia
 
Posts: 17
Joined: Tue Oct 11, 2016 10:58 am

Re: K-grid is not fully q-connected @ Q**

Postby Davide Sangalli » Wed May 31, 2017 7:59 am

Ok. Thanks.
Yeah, the setup is done reading the q-points from the s.dbph_*\

Do you still get the error "K-grid is not fully connected" when running ypp_ph after such initialization, right ?
In case I've no idea. I'll try to reproduce the problem.

Best,
D.
Davide Sangalli, PhD
CNR-ISM, Division of Ultrafast Processes in Materials (FLASHit) and MaX Centre
http://www.ism.cnr.it/en/davide-sangalli-cv/
http://www.max-centre.eu/
User avatar
Davide Sangalli
 
Posts: 316
Joined: Tue May 29, 2012 4:49 pm
Location: Via Salaria Km 29.3, CP 10, 00016, Monterotondo Stazione, Italy

Re: K-grid is not fully q-connected @ Q**

Postby elena.cannuccia » Wed May 31, 2017 8:22 am

Yes Davide, I still get the warning.

At the moment the only grid not having such a problem is a very small one, 3x3x1.
The others (6x6x2, 8x8x2, 10x10x3) are problematic.

By the way, I wonder if it is linked to the fractional translations...

Thank you for any help.

Elena
elena.cannuccia
 
Posts: 17
Joined: Tue Oct 11, 2016 10:58 am

Re: K-grid is not fully q-connected @ Q**

Postby andrea marini » Wed May 31, 2017 8:28 am

Dear Elena,

I have tried to run PW but the run is very long.

Can you, please, try to reproduce the error reducing as much as possible the parameters and making the calculation much faster?

You can heavily reduce the cut-off and/or upload somewhere (like dropbox) all pertinent files

What do you mean with "fractional translations"? Of atoms or k-points?

Andrea
Andrea MARINI
Istituto di Struttura della Materia, CNR, (Italy)
User avatar
andrea marini
 
Posts: 447
Joined: Mon Mar 16, 2009 4:27 pm

Next

Return to Initialization

Who is online

Users browsing this forum: No registered users and 1 guest

cron