VASP 5.2 installation problem
Posted: Sun Aug 21, 2011 1:55 pm
Hello everyone,
I have installed a serial version of VASP 5.2.12 on a new machine. The installation seems to have finished with no problems.
I have followed the recent recommendations given in this forum:
http://cms.mpi.univie.ac.at/vasp-forum/ ... php?2.9897
The only difference with the proposed installation above is in the line:
FC = ifort ?lsvml
The ?lsvml was necessary to avoid errors related to: libsvml.a(svml_stub_scos4.o) is referenced by DSO
After doing this change, a serial version of VASP seems to have been installed with no problems.
Running the benchmark example, VASP runs with no problems when I run the benchmark example using IALGO = 48.
However, VASP ends with Segmentation faults if I use IALGO = 38, or IALGO = 58.
The segmentation fault appears at the start of the first electronic loop:
+++++++++
POSCAR, INCAR and KPOINTS ok, starting setup
WARNING: small aliasing (wrap around) errors must be expected
FFT: planning ...( 4 )
WAVECAR not read
prediction of wavefunctions initialized
entering main loop
N E dE d eps ncg rms rms(c)
Segmentation fault (core dumped)
+++++++++
The system has the following characteristics: AMD Opteron(tm) Processor 6136 running Red Hat Enterprise 6, Intel Fortran Compiler version 11.1, GotoBlas2, FFTW 3.2.2.
The stack-size is set to unlimit.
Would you be able to provide some hints to fix this problem?
Regards,
I have installed a serial version of VASP 5.2.12 on a new machine. The installation seems to have finished with no problems.
I have followed the recent recommendations given in this forum:
http://cms.mpi.univie.ac.at/vasp-forum/ ... php?2.9897
The only difference with the proposed installation above is in the line:
FC = ifort ?lsvml
The ?lsvml was necessary to avoid errors related to: libsvml.a(svml_stub_scos4.o) is referenced by DSO
After doing this change, a serial version of VASP seems to have been installed with no problems.
Running the benchmark example, VASP runs with no problems when I run the benchmark example using IALGO = 48.
However, VASP ends with Segmentation faults if I use IALGO = 38, or IALGO = 58.
The segmentation fault appears at the start of the first electronic loop:
+++++++++
POSCAR, INCAR and KPOINTS ok, starting setup
WARNING: small aliasing (wrap around) errors must be expected
FFT: planning ...( 4 )
WAVECAR not read
prediction of wavefunctions initialized
entering main loop
N E dE d eps ncg rms rms(c)
Segmentation fault (core dumped)
+++++++++
The system has the following characteristics: AMD Opteron(tm) Processor 6136 running Red Hat Enterprise 6, Intel Fortran Compiler version 11.1, GotoBlas2, FFTW 3.2.2.
The stack-size is set to unlimit.
Would you be able to provide some hints to fix this problem?
Regards,