BSE solving stops unexpectedly

You can find here problems arising when using old releases of Yambo (< 5.0). Issues as parallelization strategy, performance issues and other technical aspects.

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

Locked
kle2797
Posts: 14
Joined: Mon Jul 12, 2021 3:21 pm

BSE solving stops unexpectedly

Post by kle2797 » Fri Apr 15, 2022 6:25 pm

Dear developers,

I am trying to solve the BSE for a spin-polarized system. I tried both slepc and diagonalization methods but I encountered different problems for each.
For Slepc, the LOG file:

Code: Select all

 <06h-15m> P1-tempo040: [SLEPC] Slower alogorithm but BSE matrix distributed over MPI tasks
P1-tempo040: [ERROR] STOP signal received while in[03.04.01] Slepc Solver @q1
P1-tempo040: [ERROR] Slepc with shells and multiple BS_mat not coded
So I guess Slepc implementation was not ready yet, and tried changing to the diagonalization method. This time, there is an error printed out:

Code: Select all

** On entry to CHEEV parameter number  5 had an illegal value
Could you help me have a look? I am not sure if this is a memory issue or from my input file. From the LOG file, I am not able to know any further. And there is only one "ndb.BS_diago_Q1" in the output folder.
I have attached my BSE input file anda LOG file for each calculation below. I hope that helps.
You do not have the required permissions to view the files attached to this post.
Last edited by kle2797 on Fri Apr 15, 2022 9:42 pm, edited 1 time in total.
Khoa Le
PhD Candidate
California Institute of Technology

Zafer Kandemir
Posts: 34
Joined: Wed May 06, 2020 9:43 am
Contact:

Re: BSE solving stops unexpectedly

Post by Zafer Kandemir » Fri Apr 15, 2022 8:58 pm

Dear Khoa,

This is the first time I've seen those errors, and I'm also wondering why those errors are occurring. I wanted to say that there are no attachments. We already calculate for a single q1-point, so there is only one file.

Best regards.
Zafer
Dr. Zafer Kandemir
Postdoctoral Researcher (Theoretical Physicist)
Department of Mechanical Engineering, Eskisehir Technical University
26555 Tepebasi-ESKISEHIR/TURKEY

kle2797
Posts: 14
Joined: Mon Jul 12, 2021 3:21 pm

Re: BSE solving stops unexpectedly

Post by kle2797 » Fri Apr 15, 2022 9:44 pm

Zafer Kandemir wrote: Fri Apr 15, 2022 8:58 pm Dear Khoa,

This is the first time I've seen those errors, and I'm also wondering why those errors are occurring. I wanted to say that there are no attachments. We already calculate for a single q1-point, so there is only one file.

Best regards.
Zafer
That was my bad. I forgot to add the files at the end... I just edited the post. Thank you.
Khoa Le
PhD Candidate
California Institute of Technology

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

Re: BSE solving stops unexpectedly

Post by Daniele Varsano » Mon Apr 18, 2022 4:29 pm

Dear Khoa,
can you also post your report file?

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/

kle2797
Posts: 14
Joined: Mon Jul 12, 2021 3:21 pm

Re: BSE solving stops unexpectedly

Post by kle2797 » Tue Apr 19, 2022 5:35 pm

Daniele Varsano wrote: Mon Apr 18, 2022 4:29 pm Dear Khoa,
can you also post your report file?

Best,
Daniele
Hello Daniele,

Yes. Here is the report file.

Best,
Khoa
You do not have the required permissions to view the files attached to this post.
Khoa Le
PhD Candidate
California Institute of Technology

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

Re: BSE solving stops unexpectedly

Post by Daniele Varsano » Wed Apr 20, 2022 8:41 am

Dear Khoa Le,

From your report it seems you have zero dimension in the antiresonant part of the BSE and this is the reason of the error you get:

Code: Select all

  [BSK] Transition groups (total):   95
  [BSK] Size (res,ares):   7455      0
  [BSK]         (total):   7455
  [BSK] Matricies      :  2
It is not clear to me anyway why this happens.
We will need to reproduce the error, can you try to reproduce the error by using minimal parameters (bands, cutoffs etc.) in order to have a fast test to look at?
If you can do that, next post QE/Yambo input files and pseudopotentials.

Many thanks,

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
Davide Sangalli
Posts: 610
Joined: Tue May 29, 2012 4:49 pm
Location: Via Salaria Km 29.3, CP 10, 00016, Monterotondo Stazione, Italy
Contact:

Re: BSE solving stops unexpectedly

Post by Davide Sangalli » Wed Apr 20, 2022 3:39 pm

Dear Khoa Le,
you are experiencing all these issues because you are trying to use the BSE to compute magnons, i.e. you are setting

Code: Select all

BSEprop="magn"
in your input file together with

Code: Select all

BSEmod="retarded"  
(or maybe "causal" in place of retarded)

In such case yambo will compute the spin-flip transitions.
It is an experimental feature and only few solvers are available.
The resonant and the anti-resonant blocks have different size and needs to be computed independently.

If you are interested in the optical properties of your material, just set

Code: Select all

BSEprop="abs"
and everything should work fine.

Best,
D.
Davide Sangalli, PhD
CNR-ISM, Division of Ultrafast Processes in Materials (FLASHit) and MaX Centre
https://sites.google.com/view/davidesangalli
http://www.max-centre.eu/

kle2797
Posts: 14
Joined: Mon Jul 12, 2021 3:21 pm

Re: BSE solving stops unexpectedly

Post by kle2797 » Wed Apr 20, 2022 10:16 pm

Davide Sangalli wrote: Wed Apr 20, 2022 3:39 pm Dear Khoa Le,
you are experiencing all these issues because you are trying to use the BSE to compute magnons, i.e. you are setting

Code: Select all

BSEprop="magn"
in your input file together with

Code: Select all

BSEmod="retarded"  
(or maybe "causal" in place of retarded)

In such case yambo will compute the spin-flip transitions.
It is an experimental feature and only few solvers are available.
The resonant and the anti-resonant blocks have different size and needs to be computed independently.

If you are interested in the optical properties of your material, just set

Code: Select all

BSEprop="abs"
and everything should work fine.

Best,
D.
Dear Davide,

Thank you for your response. I start to get it now.

Prior to this, I did the same calculation with a simpler system, using the same option

Code: Select all

BSEprop="magn"; BSEmod="retarded"
and the calculation seemed to have no problem at all. I will try to look more into it but I guess it was my particular choice of BSEbands that made the 2 blocks having the same size.

I am trying to get the wavefunction in spin-flip channel. Is there a way to compute each block seperately then?

Best,
Khoa Le
Khoa Le
PhD Candidate
California Institute of Technology

User avatar
Davide Sangalli
Posts: 610
Joined: Tue May 29, 2012 4:49 pm
Location: Via Salaria Km 29.3, CP 10, 00016, Monterotondo Stazione, Italy
Contact:

Re: BSE solving stops unexpectedly

Post by Davide Sangalli » Fri Apr 22, 2022 3:03 pm

Ok, I see.
I also suspect that the choice of bands makes the anti-resonant block of zero size.

If you specify

Code: Select all

BSEmod="resonant"  
it will compute the resonant block only.

The "anti-resonant block only" option is not coded

Best,
D.
Davide Sangalli, PhD
CNR-ISM, Division of Ultrafast Processes in Materials (FLASHit) and MaX Centre
https://sites.google.com/view/davidesangalli
http://www.max-centre.eu/

kle2797
Posts: 14
Joined: Mon Jul 12, 2021 3:21 pm

Re: BSE solving stops unexpectedly

Post by kle2797 » Mon Apr 25, 2022 2:39 am

Davide Sangalli wrote: Fri Apr 22, 2022 3:03 pm Ok, I see.
I also suspect that the choice of bands makes the anti-resonant block of zero size.

If you specify

Code: Select all

BSEmod="resonant"  
it will compute the resonant block only.

The "anti-resonant block only" option is not coded

Best,
D.
Dear Davide,

Changing to

Code: Select all

BSEmod="resonant"
solved this. I think this should be good enough for me. Thank you very much!

Best,
Khoa
Khoa Le
PhD Candidate
California Institute of Technology

Locked