eig_read=.true.

[Nandan]

Hi,

I had couple of questions about the eig_read option.

If it is set to .true.

1. The documentation shows it looks for a file ksdata.fmt with eigen values. But it keeps
giving error and is looking for seedname.eig instead. So instead of ksdata.fmt the eigen
values should be given in file name seedname.eig?

2. Since seedname.eig gives the eigen values to the final grid, does epw.x still
evaluate the eigen values to the points given in file filkf?
Because it seems that the eigen values in seedname.eig and linewidth.elself are
not identical in my case.

Thanks and regards,

Nandan.

[sponce]

Hello Nandan,

This variable is a legacy from an older version of the code and should not work at the moment.

I kept the input variable as I intend to make it work in the near future (next few months).

The idea will be to read eigenenergies on the coarse mesh (for example GW ones) and to interpolate those.

Feel free to give it a go and try to implement it now if you need it quick. It should not be too complicated as some of it is there.

If you are successful, I'm more than happy to include your improvements.

Best,

Samuel

[Nandan]

Hi Samuel,

Thanks for the reply. I will try it.

Regards,

Nandan.

[hysun]

Hi Samuel,

I have the same question (the eigen values in seedname.eig and linewidth.elself are
not identical) when I run the EPW V 4.3 .

Does the variable 'eig_read' work well in the last vertion now ?


thanks and regards,


Hysun

[sponce]

Dear Hysun,

Yes it should be working.

If you set eig_read=.true. then the code will look for a file named "seedname.eig".

Such file might contain (for example) GW eigenvalues. It has to be provided on the same grid as the nscf.in (so full homogeneous coarse grid).

Those eigenenergies will be used instead of the DFT ones.

Best wishes,
Samuel