param_read_chk: Mismatch in nntot

[hetian-chen]

This problem occurred when I calculated the band with "restart = plot", while there was no problem if I directly calculated the band. I am pretty sure that the .chk file was just generated by wannier90.x. Can anyone tell me what's wrong?

[hetian-chen]

This is what I got from w90chk2chk:

Number of bands: 50 Excluded bands: 1,2,3,4,5,6. Real lattice: read. Reciprocal lattice: read. Num kpts:729 mp_grid: read. kpt_latt: read. nntot:1906824895

[jryates]

Your second message shows us that nntot has an unphysical value. I have never seen this happen before so either: there is an issue with your compilation of wannier90 or you are doing something unusual. But that is all I can say. In order for us to investigate can you:

1. See if you can reproduce this error in a smaller test case
2. Either post those input files, or the original ones
3. Tell us which version of wannier90, and which compiler / libraries you are using.

[hetian-chen]

Your second message shows us that nntot has an unphysical value. I have never seen this happen before so either: there is an issue with your compilation of wannier90 or you are doing something unusual. But that is all I can say. In order for us to investigate can you:

1. See if you can reproduce this error in a smaller test case
2. Either post those input files, or the original ones
3. Tell us which version of wannier90, and which compiler / libraries you are using.

Thank you very much! The .chk file can be read correctly on another computer. I never met this problem on this computer before. Maybe something has been changed by others. Wannier90-3.1.0 is used and intel compilers and libraries 2016.3.210 are used. I try recompiling wannier90 but it makes no difference.

[jryates]

We don't have enough info to reproduce - and I suspect this is a local compilation issue.