Error in routine cdiagh2 (prtgkk=.true.)

[Qiaolin]

Dear EPW developers:

I am running EPW on NSCC. I try to calculate the electron-phonon vertex |g| of bulk Si, which is successful without Spin-orbital coupling (SOC). But when considering the effect of SOC, I get the following error:

===================================================================

Fermi energy is read from the input file: Ef = 6.402000 eV

===================================================================

ibndmin = 3 ebndmin = 0.456
ibndmax = 14 ebndmax = 0.647


Number of ep-matrix elements per pool : 864 ~= 6.75 Kb (@ 8 bytes/ DP)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Error in routine cdiagh2 (5):
info =/= 0
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

stopping ...


I haven't been able to find this error reported in the forum before and in the code, but it might be something obvious that I'm missing due to my lack of experience with the code. My input files: https://github.com/Isabellelin/g_matric_EPW

Best,

Qiaolin

[hlee]

Dear Qiaolin:

Your error is related to the previous reports (1. viewtopic.php?t=747 and 2. https://lists.quantum-espresso.org/pipe ... 35560.html ).

In any case, it seems that you are using the old version of EPW and QE.

Could you rerun all calculations from scratch using QE v6.7 and EPW v5.3.1 which can be downloaded at https://gitlab.com/QEF/q-e/-/releases ?

From EPW v5.3, there are several changes in the EPW input. For example, nbndskip is not used any more and the q-point list doesn't need to be added to the end of the input file. Regarding other changes, please check the Releases page at https://docs.epw-code.org/doc/Releases.html .

If you still encounter the same error with EPW v5.3.1, please provide all outputs and all inputs including GXWLGK_50.txt .

Sincerely,

H. Lee

[Qiaolin]

Dear H. Lee

Thanks for your reply.

According to your advice, I try to use the QE6.7(https://github.com/QEF/q-e/releases), but when I try to use epw.x, I encounter the problems:"forrtl: severe (59): list-directed I/O syntax error, unit -131, file ../phonons_soc/save/si.dyn_q1.xm", which did not happen in the old version QE.

All the related files are displayed in https://github.com/Isabellelin/g_matric_EPW


Sincerely,

QiaoLin

[hlee]

Dear QiaoLin:

Could you use pp.py in the EPW/bin directory instead of your pp_soc.py?

Sincerely,

H. Lee

[Qiaolin]

Dear H. Lee:

Thank you very much! With your help, I have got the electron-phonon vertex |g| with SOC.

Sincerely,

QiaoLin