runtime error with specific input in 5.4.4
Posted: Thu Nov 02, 2017 12:20 pm
Dear Vasp users and developers,
there is a strange issue with vasp5.4.4:
I compiled with ifort 14.0.5.212, mkl, gcc-4.9.2 and openmpi-2.1.2 (the same behaviour occurs with ompi-3.0.0).
A SCF calculation with the attached input INCAR-scf finished successfully. The following BANDS calculation
with the attached input INCAR-bands finished successfully with NBANDS <= 20 and gives a segfault for any
larger value of NBANDS.
Compiling the code with gcc/gfortran-4.9.2, opmpi-2.1.2 and scalapack-2.0.2 all calculations finished successfully
even with a large value of NBANDS (tested with 32).
1) Is it perhaps a bug in your code which can be fixed?
2) I would like to get a binary compiled with mkl and ifort from performance reason. Which ifort version is
recommended by the developers (I know about other issues with ifort 15 and like to avoid checking each version
myself...)?
Please let me know if you need further information.
Many thanks in advance and best regards,
Ulrike.
there is a strange issue with vasp5.4.4:
I compiled with ifort 14.0.5.212, mkl, gcc-4.9.2 and openmpi-2.1.2 (the same behaviour occurs with ompi-3.0.0).
A SCF calculation with the attached input INCAR-scf finished successfully. The following BANDS calculation
with the attached input INCAR-bands finished successfully with NBANDS <= 20 and gives a segfault for any
larger value of NBANDS.
Compiling the code with gcc/gfortran-4.9.2, opmpi-2.1.2 and scalapack-2.0.2 all calculations finished successfully
even with a large value of NBANDS (tested with 32).
1) Is it perhaps a bug in your code which can be fixed?
2) I would like to get a binary compiled with mkl and ifort from performance reason. Which ifort version is
recommended by the developers (I know about other issues with ifort 15 and like to avoid checking each version
myself...)?
Please let me know if you need further information.
Many thanks in advance and best regards,
Ulrike.