[Wien] mstar: Error: dM is not finite

[Rubel, Oleg]

dE = 0 is very strange. This should no happen. Degenerate states (or nearly degenerate |dE| < dEtol) are sorted out from the beginning and treated separately within a degenerate perturbation theory to avoid "dM = p2/dE".

What is your dEtol? (A meaningful value is 1e-4 ... 1e-6 Ry).  What happens if you raise dEtol? Do you mind sharing a link with the mommat file?

Thank you
Oleg

________________________________________
From: Wien <wien-bounces at zeus.theochem.tuwien.ac.at> on behalf of Laurence Marks <laurence.marks at gmail.com>
Sent: Sunday, December 5, 2021 14:41
To: A Mailing list for WIEN2k users
Subject: Re: [Wien] mstar: Error: dM is not finite

Obviously the line should be something like
dM = p2*dE/(dE*dE+1D-16)

---
Professor Laurence Marks
Department of Materials Science and Engineering
Northwestern University
www.numis.northwestern.edu<http://www.numis.northwestern.edu>
"Research is to see what everybody else has seen, and to think what nobody else has thought" Albert Szent-Györgyi

On Sun, Dec 5, 2021, 7:23 PM Gavin Abo <gabo13279 at gmail.com<mailto:gabo13279 at gmail.com>> wrote:

According to the README.md at [1], read_mommat_pij.f90 has to be modified for WIEN2k 19.1.  Did you do that?  If you didn't do that it might explain why your dEij(n,k) and dEij(m,k) values are both zero.

In the mstar.f90 source code (from lines 422 and 423) [2], you can see that dE needs to be non-zero to have a computed value for dM.  If dE is zero then it causes a divide by zero [3] that is an incomputable computer mathematical operation such that dM would be undefined (which is why the "Error: dM is not finite").

line 422: dE = -dEij(n,k)/2 -dEij(m,k)/2
line 423: dM = p2/dE

[1] https://github.com/rubel75/mstar<https://urldefense.com/v3/__https://github.com/rubel75/mstar__;!!Dq0X2DkFhyF93HkjWTBQKhk!A1M4TMzDqdChWXiSrMbJ6jpLdUmIOs0RXkqPT-_6ags-PpPXOBVrkJMCVKaWO_6TYYaogQ$>
[2] https://github.com/rubel75/mstar/blob/master/mstar.f90<https://urldefense.com/v3/__https://github.com/rubel75/mstar/blob/master/mstar.f90__;!!Dq0X2DkFhyF93HkjWTBQKhk!A1M4TMzDqdChWXiSrMbJ6jpLdUmIOs0RXkqPT-_6ags-PpPXOBVrkJMCVKaWO_66qkrE4g$>
[3] https://www.math.utah.edu/~pa/math/0by0.html<https://urldefense.com/v3/__https://www.math.utah.edu/*pa/math/0by0.html__;fg!!Dq0X2DkFhyF93HkjWTBQKhk!A1M4TMzDqdChWXiSrMbJ6jpLdUmIOs0RXkqPT-_6ags-PpPXOBVrkJMCVKaWO_5w-17Opg$>

On 12/5/2021 1:30 AM, Dr. K. C. Bhamu wrote:
Dear Prof. Oleg,
I am running mstar with Wien2k_19.1 with PBE+SO (20000 kpt, rkmax=8, lvns =6).
I am getting ebelow error:

 KP: 1876 bands: 856 progress:  75%
 KP: 1877 bands: 854 progress:  75%
 ikpt =                  1878
 n =                    39
 k =                    40
 m =                    39
 alpha =                     1
 beta =                     1
 pij(alpha,n,k) = (-3.6972599E-06,1.5258900E-06)
 pij(beta,k,m) = (-3.6972599E-06,-1.5258900E-06)
 pij(beta,n,k) = (-3.6972599E-06,1.5258900E-06)
 pij(alpha,k,m) = (-3.6972599E-06,-1.5258900E-06)
 dEij(n,k) =  0.0000000E+00
 dEij(m,k) =  0.0000000E+00
 dM =  (-Infinity,NaN)
 dE =   0.0000000E+00
 p2 =  (3.1996142E-11,0.0000000E+00)
Error: dM is not finite

Should I do a fresh calculation with less number of k-points or the error is about something else?
Let me know if I need to share any files.

Thanks and regards
Bhamu
_______________________________________________
Wien mailing list
Wien at zeus.theochem.tuwien.ac.at<mailto:Wien at zeus.theochem.tuwien.ac.at>
https://urldefense.com/v3/__http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien__;!!Dq0X2DkFhyF93HkjWTBQKhk!A1M4TMzDqdChWXiSrMbJ6jpLdUmIOs0RXkqPT-_6ags-PpPXOBVrkJMCVKaWO_7JRfDjLw$
SEARCH the MAILING-LIST at:  https://urldefense.com/v3/__http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html__;!!Dq0X2DkFhyF93HkjWTBQKhk!A1M4TMzDqdChWXiSrMbJ6jpLdUmIOs0RXkqPT-_6ags-PpPXOBVrkJMCVKaWO_6th4E_9Q$