2D kpoint grid being interpreted as 3D

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.

Moderators: Davide Sangalli, andrea.ferretti, myrta gruning, Daniele Varsano

Post Reply
arb83@cam.ac.uk
Posts: 98
Joined: Thu Jul 02, 2020 3:56 pm

Re: 2D kpoint grid being interpreted as 3D

Post by arb83@cam.ac.uk » Tue Dec 01, 2020 7:32 pm

Dear Daniele

I've just noticed that in the static screening calculation I didn't include the line 'NoDiagSC' and the k-point grid was interpreted as 5 6 1 instead of 6 6 1. I'm going back and re-running that step with the line 'NoDiagSC' included. Hopefully this will resolve issues further down the line.

Best wishes,

Alan
Alan Bowman
University of Cambridge

arb83@cam.ac.uk
Posts: 98
Joined: Thu Jul 02, 2020 3:56 pm

Re: 2D kpoint grid being interpreted as 3D

Post by arb83@cam.ac.uk » Wed Dec 02, 2020 6:46 pm

Dear Daniele

Unfortunately I'm still observing the same problem - the BSE static dielectric screening says that the X grid is uniform, while the BSE kernel and solver says the X grid is not uniform. What should I do to try and resolve this? If you'd like the new input files (fairly certain they're same as the other ones I uploaded) let me know.

Best wishes,

Alan
Alan Bowman
University of Cambridge

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: 2D kpoint grid being interpreted as 3D

Post by Daniele Varsano » Wed Dec 02, 2020 7:31 pm

Dear Alan,
the static screening and the BSE need different checks about the uniformity of the grid.
In the previous report you post, the ndb.kindx was read from a previous calculation and the grid was not uniform:

Code: Select all

[RD./SAVE//ndb.kindx]---------------------------------------
  Fragmentation                    :no
  Polarization last K   : 20
  QP states             :  1  20
  X grid is uniform     :no
  BS scattering         :no
  COLL scattering       :no
The NoDiagSC was not read in the input.
You can try to delete the ndb.kindx, and redo the calculation including the NoDiagSC.
Be sure that the variable is correctly read looking at the mirror at the end of the report. If it fails you are in a situation where it is impossible to Yambo to see the grid as uniform.

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/

arb83@cam.ac.uk
Posts: 98
Joined: Thu Jul 02, 2020 3:56 pm

Re: 2D kpoint grid being interpreted as 3D

Post by arb83@cam.ac.uk » Thu Dec 03, 2020 3:38 pm

Dear Daniele

This worked (i.e. deleting the ndb.kindx), thank you hugely.

One final question - I'm trying to get ypp to calculate the exciton energies. It is interpreting the kpoint grid as 561 instead of 661 - does ypp work with NoDiagSC? And if so, how do I view an input file for the option '-e s 1'? I tried a simple '-F' approach but this didn't seem to work.

Best wishes,

Alan
Alan Bowman
University of Cambridge

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: 2D kpoint grid being interpreted as 3D

Post by Daniele Varsano » Thu Dec 03, 2020 3:51 pm

Dear Alan,

ypp -e s 1 , it reads eigenvalues and residue contained in the BSE database and sorts them according to energy or strength, it does not need any input. It does not recalculate the k point grid and should provide directly the output files.

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/

arb83@cam.ac.uk
Posts: 98
Joined: Thu Jul 02, 2020 3:56 pm

Re: 2D kpoint grid being interpreted as 3D

Post by arb83@cam.ac.uk » Thu Dec 03, 2020 6:14 pm

dear Daniele

I'm slightly confused, in the BSE calculation run file I have:

"
[02.04] K-grid lattice
======================

Using the new bz sampling setup
Smallest q vectors (crystal coordinates) :
0.166667 0.166667 1.000000
Compatible Grid is 3D
B1 [rlu]=-.1667 0.4264E-8 -.5960E-7
B2 =0.2765E-8 0.1667 0.7451E-8
B3 =0.5000 0.1342E-7 -1.000
Grid dimensions : 6 6 1
K lattice UC volume [au]:0.6722E-3
"

but when I run "ypp -e s 1" I receive the output:

"
<---> [01] CPU structure, Files & I/O Directories
<---> [02] Y(ambo) P(ost)/(re) P(rocessor)
<---> [03] Core DB
<---> :: Electrons : 416.0000
<---> :: Temperature [ev]: 0.000000
<---> :: Lattice factors [a.u.]: 14.51095 11.38441 62.33173
<---> :: K points : 20
<---> :: Bands : 1800
<---> :: Symmetries : 2
<---> :: RL vectors : 489531
<---> [04] K-point grid
<---> :: Q-points (IBZ): 20
<---> :: X K-points (IBZ): 20
<---> [05] CORE Variables Setup
<---> [05.01] Unit cells
<---> [05.02] Symmetries
<---> [05.03] RL shells
<---> [05.04] K-grid lattice
<---> Grid dimensions : 5 6 1
<---> [05.05] Energies [ev] & Occupations
<03s> [06] Excitonic Properties
<03s> [07] Reporting sorted Energies and Strengths
<04s> :: Sorting energies
<04s> [08] Game Over & Game summary
"

i.e. the Grid dimensions are different in the two cases. I expect I don't need to be concerned about this (as ypp is not interpreting values with NoDiagSC) but if you could confirm it'd be fantastic.

Best wishes,

Alan
Alan Bowman
University of Cambridge

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: 2D kpoint grid being interpreted as 3D

Post by Daniele Varsano » Fri Dec 04, 2020 6:26 am

Dear Alan,
yes, it is as you say, and here it is harmless.

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/

Post Reply