seg fault after install vasp 5.3.5
Posted: Wed Jun 24, 2015 6:49 pm
Dear Vasp Admins and Users,
I have been able to compile many different flavors of vasp 5.3.5 e.g. vasp, vasp-gamma, intel-12.1.3, intel-14.0.2, serial, parallel. Yet all of the binaries fail and end with a segmentation fault after entering the main loop. The test file is just a CO2 molecule and runs successfully with an older vasp 4.6 binary. Debugging shows the interrupt occurring in the hamil_mp_hamiltmu_ () function. The makefile is very similar to the linux_ifc_P4. I am using mpif90 from an installed opempi-1.6.3 (compiled using intel-12.1.3) and ifort from the same intel install. The FFTW library was from the mkl/interfaces/fftw. Instead BLAS= -lguide -lmkl, I used BLAS= -lpthread -lmkl. I have tried the same makefile on two separate clusters and the problem is the same. Please advise on how to fix this problem. Attached are the INCAR and POSCAR files and the Makefile (truncated after general rules as I didn't change anything).
Thanks for your help,
Kevin
I have been able to compile many different flavors of vasp 5.3.5 e.g. vasp, vasp-gamma, intel-12.1.3, intel-14.0.2, serial, parallel. Yet all of the binaries fail and end with a segmentation fault after entering the main loop. The test file is just a CO2 molecule and runs successfully with an older vasp 4.6 binary. Debugging shows the interrupt occurring in the hamil_mp_hamiltmu_ () function. The makefile is very similar to the linux_ifc_P4. I am using mpif90 from an installed opempi-1.6.3 (compiled using intel-12.1.3) and ifort from the same intel install. The FFTW library was from the mkl/interfaces/fftw. Instead BLAS= -lguide -lmkl, I used BLAS= -lpthread -lmkl. I have tried the same makefile on two separate clusters and the problem is the same. Please advise on how to fix this problem. Attached are the INCAR and POSCAR files and the Makefile (truncated after general rules as I didn't change anything).
Thanks for your help,
Kevin