Added logic to angle code to avoid singularity; improved tabulated bonded kernel arguments; commented out (non-working) get_energy code

Commented out cudaSetDevice statements (these were added prematurely and should go through GPUManager)

computeTabulateBonds works, bug computing forces fixed, bondList is created, but not implemented yet.

Conflicts:
	ComputeForce.cuh
	TabulatedPotential.h

Code works now for small systems; probably fails to allocate a single large chunk of mem for pairlists

computeTabulatedKernel now runs in only 11 microseconds (22500 particles); pairlist creation kernel is quite slow (65 ms) and crashes on update