EPW Forum

Dear experts and users,

I want to plot the wannier bands using band_plot = .true. flag.
I have already made the isotropic Eliashberg calculation so everything should be ready for calculating the wannier bands too.
So, I modified the input:

ep_coupling = .true.
elph = .true.
kmaps = .true.
epbwrite = .false.
epbread = .false.
epwwrite = .false.
epwread = .true.

dvscf_dir = '../phonons/save'
band_plot = .true.
!a2f = .false.
!phonselfen = .true.
!delta_approx = .true.
filqf = 'myprefix_band.kpt'
filkf = 'myprefix_band.kpt'
!mp_mesh_k = .true.

As I run the calculation, I get an error:
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Error in routine loadqmesh_serial (1):
ERROR: Specify either crystal or cartesian coordinates in the filqf file
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Which is odd, since EPW created the 'myprefix_band.kpt' (filqf) file with the crystal coordinates in it, and the file certainly does exist with correct contents. In order to create it I used:
wdata(2) = 'bands_plot = true'
wdata(3) = ' begin kpoint_path'
wdata(4) = 'G 0.000000000 0.000000000 0.000000000 M 0.500000000 0.000000000 0.000000000'
wdata(5) = ' M 0.500000000 0.000000000 0.000000000 K 0.333333333 0.333333333 0.000000000'
wdata(6) = ' K 0.333333333 0.333333333 0.000000000 G 0.000000000 0.000000000 0.000000000'
wdata(7) = ' end kpoint_path'

Could you please tell me what is wrong then?
I get the problem when I use EPW 5.2.0.
With EPW 5.1.0 I calculated bands successfully many times, but I set both epbread and epwread to .true. simultaneously in the input, which is not allowed in the 5.2.0 version. Could it be the problem? Could you please tell me what is the correct procedure for EPW 5.2.0?

UPDATE:
I just checked, and the problem seem to be that restart option doesn't work properly in my version 5.2.0. I can't even reproduce my already obtained isotropic Eliashberg results. When I run the restart calculation for isotropic Eliashberg, i get:
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Error in routine invmat (1):
error in DGETRF
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Curiously, if I load version 5.1.0 in the same directory and set both epbread and epwread to .true. simultaneously in the input, none of the above mentioned problems occur.

Thank you.
Best,
Mikhail

Dear Mikhail:

Could you provide me with the file of myprefix_band.kpt?

Regarding the restart problem, I don't know much about EPW v5.1, but there might be backward compatibility issues.

Sincerely,

H. Lee

Thank you for the quick reply.
The file doesn't want to attach for some reason.
It looks exactly similar to the files that the version 5.1.0 generates.
I paste the several first lines of it below:

Dear Mikhail:

Could you replace the first line by the following?

(if your coordinates are in crystal units) or

(if your coordinates are in Cartesian units) Sincerely,

H. Lee

Dear Mr. Lee,

Your solution seems to have worked. Thank you very much !
I am a bit puzzled now why my restart calculation crashes when I want to calculate isotropic/anisotropic Eliashberg.
I made the first run which included isotropic Eliashberg, and it worked very well. But now when I tried to calculate it again using the restart option, I got this QE error:
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Error in routine invmat (1):
error in DGETRF
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

I don't believe that there is anything wrong with my input since I managed to calculate the isotropic Eliashberg function.
Could you maybe suggest something please?

Best,
Mikhail

Dear Mikhail:

Regarding the restart problem, I don't know much about EPW v5.1, but there might be backward compatibility issues.

I am still suspecting that this error might come from the backward compatibility issues.
I don't know about EPW v5.1, but there might be some changes in the file structures for restart.

So if you are using the files from EPW v5.1 in restarting with EPW v5.2 and there are some changes in the file structures, you would encounter the error; the error you encountered means that there is something wrong in the arguments for matrix inversion.

Sincerely,

H. Lee

Dear Mr. Lee,

I must have given a wrong impression.
I did everything within the 5.2.0 version.
If I try to restart with 5.2.0, I immediately (I mean within a second, so in the very beginning) get an error:
Finish reading .egnv file
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Error in routine invmat (1):
error in DGETRF
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
If I try to restart with 5.1.0 (just as a test), and don't get this particular error, but the code typically crashes afterwards exactly due to backward compatibility issues.

Best,
Mikhail

Dear Mikhail:

I guess that in EPW v5.2 there is some issue with reading some basic inputs such as direct and reciprocal lattice parameters; probably, when restarting superconductivity calculations, EPW seems not to read them.

You can confirm it by modifying the code to print the arrays of at and bg.

Sincerely,

H. Lee

Dear Mr. Lee,

Thank you for your response!
I didn't quite get what I should confirm and how. Could you please elaborate?
Is this then an issue with the version of the code or is it a bad installation or something?
With the version 5.1 I completed the calculation without any problems.

Best,
Mikhail

Dear Mikhail:

I guess that in EPW v5.2 there is some issue with reading some basic inputs such as direct and reciprocal lattice parameters; probably, when restarting superconductivity calculations, EPW seems not to read them.

You can confirm it by modifying the code to print the arrays of at and bg.

You told me that you immediately encounter the error after printing "Finish reading .egnv file".
This means that there is something wrong in the subroutine of read_kqmap in io_eliashberg.f90.

In this subroutine, there is a call to set_sym_bl if mp_mesh_k=T (I guess this is your case).
As I guessed, if EPW doesn't import at and bg when restarting, you encounter the error in invmat.

So first you need to confirm it by adding the line of "USE cell_base, ONLY : at, bg" in the header of subroutine of read_kqmap and also adding the line of "WRITE(stdout, *) at, bg" in the suitable place of the same subroutine.

Sincerely,

H. Lee