Sub-optimal issues from aiida-quantumespresso

[unkcpz]

• Too many k-points in PDOS calculation aiidateam/aiida-quantumespresso#893
• PDOSWorkChain - align energy range to fermi level aiidateam/aiida-quantumespresso#764
• Change nbnd logic for QeAppWorkchain #362
• PwRelaxWorkChain: Revisit meta iterations and final SCF aiidateam/aiida-quantumespresso#705

@mbercx I create this issue for collecting the issues that need to be solved from aiida-quantumespresso. So we can have it tracked in for the next stable release v23.10.0

[mbercx]

Phew, the pressure is on! 😅

[mbercx]

Hi @unkcpz! A couple of notes:

1. Too many k-points in PDOS calculation aiidateam/aiida-quantumespresso#893

As described in the issue, this should in my view only be resolved by either:

• Increasing the precise k-points distance for the NSCF in the PDOS protocol
• Compiling Quantum ESPRESSO in the QEapp to handle more k-points.

Since the choice for precise here is rather arbitrary, I'm happy to change it to e.g. 0.8, but that still won't guarantee that you don't have to many k-points for some structures.

2. PDOSWorkChain - align energy range to fermi level aiidateam/aiida-quantumespresso#764

This one is fair, I should probably review @t-reents' PR soon. I'll prioritise for the next week! ^^

3. Change nbnd logic for QeAppWorkchain #362

Isn't this an issue on the QEapp? It's true that reducing the nbands_factor in the protocol is probably too high, which I'm happy to reduce. Is this why the issue is listed here?

4. PwRelaxWorkChain: Revisit meta iterations and final SCF aiidateam/aiida-quantumespresso#705

Although the new approach is indeed slightly better, the "old" PwRelaxWorkChain does a fine job in most cases. No need to block a publicly advertised release for this. That said, I should probably finalise the corresponding PR anyways.