NaN of quasiparticle energies with v-5.2.0

Concerns issues with computing quasiparticle corrections to the DFT eigenvalues - i.e., the self-energy within the GW approximation (-g n), or considering the Hartree-Fock exchange only (-x)

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

Post Reply
Online
sdwang
Posts: 287
Joined: Fri Apr 09, 2010 12:30 pm

NaN of quasiparticle energies with v-5.2.0

Post by sdwang » Thu Aug 03, 2023 4:37 am

Dear developers,
I got the GW results recently using Yambo (v-5.2.0), and all the quasiparticle energies are NaN in the output. Attached are the report, log, and output files. (I never met this before:? , and it seems that the HF energies are normal)
Thanks!

Shudong
You do not have the required permissions to view the files attached to this post.
S. D. Wang
IMU,HOHHOT,CHINA
E-mail: sdwang@imu.edu.cn

User avatar
Daniele Varsano
Posts: 3816
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: NaN of quasiparticle energies with v-5.2.0

Post by Daniele Varsano » Thu Aug 24, 2023 9:59 am

Dear Shudong,
thanks for reporting.

Can you check if NaN already appear in the ndb.pp_fragments* files, or they are produced afterwards?
You can inspect ndb.pp_fragments* files by using the ncdump command of the netcdf library, or reading the netcdf file by using python scripting.

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/

Online
sdwang
Posts: 287
Joined: Fri Apr 09, 2010 12:30 pm

Re: NaN of quasiparticle energies with v-5.2.0

Post by sdwang » Thu Aug 24, 2023 10:13 am

Dear Daniele,
I dumped the ndb.pp_fregment as below:

Code: Select all

netcdf ndb {
dimensions:
        D_0000000001 = 1 ;
        D_0000000002 = 2 ;
        D_0000002079 = 2079 ;
variables:
        float FREQ_PARS_sec_iq32(D_0000000001) ;
        float FREQ_sec_iq32(D_0000000002, D_0000000002) ;
        float X_Q_32(D_0000000002, D_0000002079, D_0000002079, D_0000000002) ;

// global attributes:
                :_NCProperties = "version=2,netcdf=4.9.0,hdf5=1.12.2" ;
data:

 FREQ_PARS_sec_iq32 = 0.9999999 ;

 FREQ_sec_iq32 =
  1.192093e-07, 0,
  0, 0.9999999 ;

 X_Q_32 =
  -0.3714867, 3.151352e-08,
  0.1860252, 0.1691717,
  0.1867038, -0.1691292,
  0.006388897, -0.0385413,
  0.005320549, 0.03745158,
...
It seems OK.

Best

Shudong
S. D. Wang
IMU,HOHHOT,CHINA
E-mail: sdwang@imu.edu.cn

User avatar
Daniele Varsano
Posts: 3816
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: NaN of quasiparticle energies with v-5.2.0

Post by Daniele Varsano » Mon Aug 28, 2023 10:03 am

Dear Shudong,
not easy to spot the problem. I can see that this is massive calculations (several days).
Can you try to repeat the calculation for the QP only at Gamma point? It should last about 4 hours and let's see if the problem is
reproducible.

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