Can GW for system which appears to be metallic be accurate?

[askhetan]

Dear VASP users and develoopers,

I finished GW calculations for a slab-system (using VASP) which appears to have band gap at every k point but the LUMO is lower than HOMO when considered overall. At the end of the log file I get:

CHECK_OCCUPATIONS: ERROR, system is not insulating.

I do not know the theory of GW calculations in and out but I want to understand that if the system is metallic, is there any use in doing GW simulations ? Are the results that I have now completely crap because the system is not insulating ? Can I get atleast the correct band structure ?

[askhetan]

Actually, I have managed to converge it without errors using the recommendation from the website. I used IMIX =0 with TIME = 0.1 (instead of the default 0.4) and now the system is still metallic but doesn't give any occupation errors.