[DFTB-Plus-User] MPI Version crashes
balint.aradi at bccms.uni-bremen.de
Tue Sep 3 13:36:47 CEST 2013
> After correcting a few issues, such as array allocationswhich did not go
> through, at least with the GCCcompilers suite, I managed to obtain
Well, we never tested the GCC suite as it apparently can't cope with
some Fortran 2003 features we have now in the code...
> but it crashes nearly immediatelyupon execution. I tried to locate the
> source of the problem, and it seems that some structures are used
> without being properly initialised.
Hard to say. The dummy arguments in initgrids() where your log indicates
the address 0x0 are OPTIONAL arguments, which may or may not be present
when the routine are called. So, those not having a proper memory
address should not make the code crash.
Could you please try to compile the code with the ifort 12.1 compiler or
a recent NAG compiler? They are the one which shouldn't have problems
with it. Alternatively, if you do not have access to those compilers,
you can send me a self contained input, then at least I can check what
our version of the code does with it.
Ps. Actually, the released code indeed has a bug (I'll put on the new
sources today), but that occurs only during MD simulations at the first
geometry step if at all.
Dr. Bálint Aradi
Bremen Center for Computational Materials Science, University of Bremen
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 261 bytes
Desc: OpenPGP digital signature
More information about the DFTB-Plus-User