The 598 zombie processes are the problem. If you do a "ps axf" you should see the parent of those processes, which may give a clue, as the parent is not cleaning up properly.
Check the walk status file(s) /usr/local/morph3/texis/2016Rules/db*.long and the vortex.log to see if anything illustrative is logged about the profile startup.
Also, what's the memory situation, reported by "free", of the machine before starting the walk?
ps auxf will provide more info about parents and process sizes etc.