Infinite loop in H.fermi_level

In my ignorance, I had a Hamiltonian and I wanted to calculate its fermi level. I thought, well that’s easy, just H.fermi_level(). But then my laptop started to heat up and hyperventilate. And I was like, man, this must be an expensive calculation.

However, it was just that the _Ef function was running infinitely because min_Ef and max_Ef were the same. This happened because q (which I initially thought was the net charge) needs to be set to the number of electrons in the system. Since I was not providing a q, it was taken from my geometry q0 attribute, which was set to 0.

I have two questions:

• If q is the number of electrons, does it make sense to check beforehand if the user is asking for something impossible? And specially if it’s 0, assume that is the sum of all atoms Z?
• Wouldn’t it be more useful that q was the net charge? So that the user does not need to know how many electrons are there in their system.

edit: I realised that what I was proposing does not make sense because you are not taking into consideration all electrons necessarily, sorry. Maybe the check that I proposed in the next comment is useful?

[zerothi]

Ok, I see now why. I have added a new commit which fixes it 😉

Thanks!

[pfebrer]

I don’t think this makes sense, it seems complicated and not necessary. You should change the atoms with a charge to control this.

Ah ok, I didn’t know you could do that 👍

How many k-points are you using?

The default, which is just the Gamma point, right?