probable error in gw_si.py yambo-py scrpit

Post here any question you encounter when running the scripts of the yambo-py suite. Post here problem strictly to the python interface as problem coming from the yambo runs should go in the appropriate subforum.

Moderators: amolina, miranda.henrique

probable error in gw_si.py yambo-py scrpit

Postby sitangshu » Tue Jan 10, 2017 6:05 pm

Dear Sir,



I am trying to plot the GW band structure using first the ground state solution from a bulk relaxed Si by using the script "gs_si.py". (I am repeating the example in the yambo-py documentation). This is an aboslutely nice file to give the ground state bandstructure, whcih I obtained effortlessly. I have generated all the fully relaxed nscf, scf, and bands files and folders with scf and nscf k_points as 12, 12, 12 respectively and dg_kpoints as 14, 14, 14. This resulted in a k00001-k00189 number of bands in the "bands" folder with xml data.



Now, in order to get the GW band structure, I started with the [recently (today's) updated yambo-py files and scripts] plasmon pole model given in the script "gw_si.py". While running the file, I am getting an error about "Dynamical Dielectric Matrix (ppa): NetCDF not a directory". I checked this back in Yambo and found out (also from Daniele's reply from my last post in the "self energy" section forum today) and from viewtopic.php?f=14&t=252 some trick to reduce the memory.

Can you please help me out where should I write this -S?



With regards,

Sitangshu
You do not have the required permissions to view the files attached to this post.
Sitangshu Bhattacharya
Indian Institute of Information Technology-Allahabad
India
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
sitangshu
 
Posts: 93
Joined: Thu Jan 05, 2017 8:08 am

Re: probable error in gw_si.py yambo-py scrpit

Postby amolina » Wed Jan 11, 2017 4:48 pm

Dear Sitangshu,

I cannot reproduce this error but it seems the netCDF files.

Try to remove the gw and database folder and start the simulations from the scratch. It should work.

Cheers,

Alejandro.
Alejandro Molina-Sánchez
Physics and Materials Research Unit
Campus Limpertsberg
Université du Luxembourg
User avatar
amolina
 
Posts: 111
Joined: Fri Jul 15, 2011 11:23 am
Location: Luxembourg

Re: probable error in gw_si.py yambo-py scrpit

Postby sitangshu » Thu Jan 12, 2017 1:25 pm

Dear Alejandro,

I am running the script gw_modify_si.py. It is running smoothly!
Now I am struggling to get the correct way to generate a converged and self-consistent GW (and not G0W0/GW0) band structure :)

Let me see, ...

With regards and thanks!

Sitangshu
Sitangshu Bhattacharya
Indian Institute of Information Technology-Allahabad
India
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
sitangshu
 
Posts: 93
Joined: Thu Jan 05, 2017 8:08 am

Re: probable error in gw_si.py yambo-py scrpit

Postby sitangshu » Sat Jan 14, 2017 7:48 am

Dear Alejandro,

So, now I have completed the GW analysis with a plasmon pole approximation using a 10x10x10 nscf bulk Si DFT run data. I have attached the r-reference and the o-reference.qp file. Some couple of confusion is still with me:
1. I inspect the r-file. I see that there ARE the "External QP correction (G)" added but no ""External QP correction (X)". Does this means that I ran a GW0 analysis? I mentioned in the run level as: -d -k hartree -p p -g n -V all. If so, how do I do a self-consistent GW.
2. I understand that the run obtained the dynamic screening part of the dielecric function. The QP corrections are done for bands 2 to 4 for a total 1-116 k points.
3. Section 6.01 reports that the HF occupation with Indirect gaps and direct gaps. What does the two vales in each line means.
4. When I inspect the .qp file, I see there are E0 (KSdata) and E-E0 (GW correction) columns. To plot the band structure, is this sufficient? Ofcourse, I think I need to arrange the data in proper format. Should I plot Col. 1 vs Col. 3 for GW band structure?
5. Although, now I am struggling to write a script for arranging these data, can you please tell if these set of data are sufficient to give a smoother band structure, or I still need to increase the k-point nscf for this run?

With regards,
Sitangshu
You do not have the required permissions to view the files attached to this post.
Sitangshu Bhattacharya
Indian Institute of Information Technology-Allahabad
India
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
sitangshu
 
Posts: 93
Joined: Thu Jan 05, 2017 8:08 am

Re: probable error in gw_si.py yambo-py scrpit

Postby Daniele Varsano » Wed Feb 01, 2017 10:16 am

Dear Sitangshu

1. I inspect the r-file. I see that there ARE the "External QP correction (G)" added but no ""External QP correction (X)". Does this means that I ran a GW0 analysis? I mentioned in the run level as: -d -k hartree -p p -g n -V all.


It means that energies have been corrected but not inserted in the screening. It is the standard G0W0.

If so, how do I do a self-consistent GW.


You can do the so called self consistent evGW (eigenvalues only) here you can find a step by step guide for doing that by Dr. Attaccalite:
http://www.attaccalite.com/self-consist ... ambo-code/

2. I understand that the run obtained the dynamic screening part of the dielecric function. The QP corrections are done for bands 2 to 4 for a total 1-116 k points.

Correct
3. Section 6.01 reports that the HF occupation with Indirect gaps and direct gaps. What does the two vales in each line means.


The first number is indeed the gap, the second one should be the max gap involving the two frontiers bands (max direct and max indirect), I'm not 100% sure of that but you can verify inspecting your calculated bands.

4. When I inspect the .qp file, I see there are E0 (KSdata) and E-E0 (GW correction) columns. To plot the band structure, is this sufficient? Ofcourse, I think I need to arrange the data in proper format. Should I plot Col. 1 vs Col. 3 for GW band structure?



5. Although, now I am struggling to write a script for arranging these data, can you please tell if these set of data are sufficient to give a smoother band structure, or I still need to increase the k-point nscf for this run?


if you plot col 1 and col2 you have KS energies, for the GW just plot col 1 and col.3+Col.4.
OF course these are the GW energies calculated for the bands you indicated according to your sampling. If you want a band structure along a defined path crossing high symmetry points you need to interpolate them: there is a tool in ypp (yambo post processing) to do that (http://www.yambo-code.org/input_file/ypp/index.php)

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


Return to Yambo-py

Who is online

Users browsing this forum: No registered users and 1 guest

cron