error in N pseudopotential

[guoxingmiao]

We run into a strange problem with VASP on one particular pseudopotential. When I use the PBE54 version of N pseudopotential, on a box running CentOS 7, I got an error message as attached. This happens whenever this N element is included in any supercell.

To narrow it down, if I use a different N pseudopotential, such as the N_s or N_h PBE54 versions, no such error. To rule out the possibility of file corruption, I downloaded the potpaw again and same error persists. However, when I try to run exactly the same files on a different box running Ubuntu, no such error.

Do you know what could be the problem? For the time being we only had the problem with N, all other elements we used worked well.

[martin.schlipf]

Can you please post the INCAR, KPOINTS and POSCAR file, so that we can reproduce the problem.

I also could not open the image you attached, could you copy it in textual form, please?

[guoxingmiao]

Here is the error message in jpg format, tif did not work.
How to upload things like INCAR in here, or just paste texts?

Untitled-1 copy.jpg

[martin.schlipf]

When you create a new forum post there is the edit area, where you type the text. Below are the three buttons (save draft, preview, submit). If you scroll down a bit further, you will see two tabs, options and attachments, click on the latter and "add files" to upload files.
Of course if your input files are short, you could also just post them in the thread.

The error message you posted looks almost like a bug in gfortran, so could you also tell me the exact version of gfortran you are using.

[guoxingmiao]

The complier is: GNU Fortran (GCC) 4.8.5 20150623 (Red Hat 4.8.5-44)
Files cannot be attached on this forum because they have wrong(no) extension. But they should not be the problem as the same files run well on a different server. I removed all other elemets except N and the error persists.

[martin.schlipf]

Okay, the issue here seems that it is just a very old version of the compiler (6 years). I would suggest to upgrade to a newer version of gcc/gfortran then the problem will likely disappear.