Page 1 of 1

odd error 2016.2

Posted: 28 Jun 2017, 11:27
by michaljz
hello,

I have installed and I am beginning to use DALTON2016.2 on a new computer.
There is a problem with a relatively small job, please find attached
the *dal, *.mol files and the output.

It is a small basis set DFT calculation of spin-spin coupling constants
in a medium size molecule, and the problem is odd because:

a) the job runs fine when I eliminate the DSO term,
b) the job runs fine when I eliminate symmetry! (forbid it in mol file)
- although a priori it becomes larger and more complicated
c) the job runs fine on two other computers where I have 2013 and 2015
Dalton versions.

The error is always "Segmentation fault", always in or immediately after
the DSO term, and might be related to the linking of system
libraries (libc-2.22.so... etc).

However, we have tried a dozen or more combinations, scalar and parallel,
with Intel MPI and OpenMPI; most of them pass succesfully ALL the provided
Dalton tests, habdle this calculation without symmetry, but not a single
compiled version managed to complete this calculation with symmetry.

All suggestions will be welcome.

Michal (Jaszunski)