Crash in symetry detection
Posted: Fri Mar 25, 2016 8:25 pm
Hello,
A user of our cluster has not been able to run a 4x4x1 Sn supercell containing a total of 128 atoms on the latest patched version of VASP 5.4.1.
After tracking down the bug, it looks like a segmentation fault occurs in routine FSYM (symlib.F) called during "Analysis of constrained symmetry for selective dynamics".
The bug does not occur if one breaks the symmetry of the supercell by slightly modifying one atomic position or when deactivating symmetry detection with "ISYM=0".
The bug has been reproduced with several versions of the Intel compiler (14, 15 and 16).
Best regards,
Jaime
A user of our cluster has not been able to run a 4x4x1 Sn supercell containing a total of 128 atoms on the latest patched version of VASP 5.4.1.
After tracking down the bug, it looks like a segmentation fault occurs in routine FSYM (symlib.F) called during "Analysis of constrained symmetry for selective dynamics".
The bug does not occur if one breaks the symmetry of the supercell by slightly modifying one atomic position or when deactivating symmetry detection with "ISYM=0".
The bug has been reproduced with several versions of the Intel compiler (14, 15 and 16).
Best regards,
Jaime