Issue with using vasp 5.2

Message

zhigangmei · #1 Post by **zhigangmei** » Tue Apr 24, 2012 10:04 pm

Hi,

I have compiled vasp 5.2.12 without no problem. There is also no problem to do regular calculations such as structure relaxation and static total energy calculations.

Problems arise when I do linear response calculations, such as dielectric constants, Born effective charges and force constant. If there is no NPAR set in INCAR, there is warning suggesting to set NPAR in standard out file, and calculation itself never starts, with error message in standard error, "mpirun noticed that process rank 23 with PID 3168 on node n25 exited on signal 11 (Segmentation fault)."

While I set NPAR in INCAR, calculation do start, but stops somewhere stating "VASP internal routines have requested a change of the k-point set. Unfortunately this is only possible if NPAR=number of nodes. Please remove the tag NPAR from the INCAR file and restart the calculations."

I am really confused what to do right now. I am not sure if this is related to my compilation of vasp 5, which is compiled using openmpi 1.4.5, MKL blas and lapack.

If anyone had similar problem, please let me know how to solve it.

Thanks,
Zhi-Gang

zhigangmei · #2 Post by **zhigangmei** » Wed Apr 25, 2012 4:12 am

There is another thing I forgot to mention that if only single node is used, there will be no above mentioned problem. However, it would become a huge limitation if I want to study relative large systems.

Zhi-Gang

zhigangmei · #3 Post by **zhigangmei** » Wed Apr 25, 2012 5:27 pm

To update:
Problems seems solved by changing to mpich2 compiler. Hope this may help those who have similar issues with compiling vasp 5.2.

Zhi-Gang