Changed how crossover dihedrals are written when local_twist is False; fixed routines that walk through helices

Streamlined walking through strands for atomic model; added on_fwd_strand attribute to Location class

Fixed some small bugs; updated intrahelical iteration to use linked-list scheme; updated bead generation

Updated run script to remove unneeded simulation steps and make fine-resolution model run longer; also added dependency on pynvml to query gpus

Made atomic simulation box smaller, because it's usually too big, causing namd to crash; TODO: calculate size from object

Conflicts:
	atomicModel.py