Hi Axel, I nailed the segmentation fault down to the Mergesort
routine 'do_merge(...)' but cannot figure why at some point befor
an array address (?) gets inserted into the 'buf[]'-list instead of the
atom id of the last atom inserted in the MC step. I'll add an
overview-image of call-stack and variables at the point of the segfault,
maybe you can figure something out.
BTW: when sorting the output, the "real" atom sequence seems to get
altered too leading to a different trajectory in the Hybrid MC/MD run
(hopefully leading to the same thermodynamics). I'll add another image
to this post (number of atoms via mc/md step). Mergesort/Qsort lead to
identical trajectories (until the system segfaults eventually with
Mergesort).
=>
[snapshot-dbg-sort.png]
[wc-spc-shake-ppf.png]
