Merge branch 'master' into collected-small-changes

   \left[ 1 - \frac{\left(5 ~ a^6+45~a^4~r^2+63~a^2~r^4+15~r^6\right) \sigma^6}

   {15 \left(a-r\right)^6 \left( a+r \right)^6} \right], \quad r < r_c

where :math:A_{cs}` is the Hamaker constant, *a* is the radius of the colloidal

where :math:`A_{cs}` is the Hamaker constant, *a* is the radius of the colloidal

particle, and :math:`r_c` is the cutoff.  This formula is derived from the

colloid-colloid interaction, letting one of the particle sizes go to

zero.

location of the (rightmost) minimum of the potential, as explained in

the syntax section above.

This potential was first used in (Cooke)_#CKD for a coarse-grained lipid

This potential was first used in :ref:`(Cooke) <CKD>` for a coarse-grained lipid

membrane model.  It is generally very useful as a non-specific

interaction potential because it is fully adjustable in depth and width

while joining the minimum at (sigma, -epsilon) and zero at (cutoff, 0)

than *sigma*\ .

If the *wca* option is used then a Weeks-Chandler-Andersen potential

(Weeks)_#WCA is added to the above specified cosine-squared potential,

:ref:`(Weeks) <WCA>` is added to the above specified cosine-squared potential,

specifically the following:

.. math::

   pair_coeff * * 1.0 1.0

   pair_coeff 1 1 1.1 2.8 3.0 3.2

   pair_style buck 2.5 3.0

   pair_style buck/mdf 2.5 3.0

   pair_coeff * * 100.0 1.5 200.0

   pair_coeff * * 100.0 1.5 200.0 3.0 3.5

namespace LAMMPS_NS {

template class AngleCharmmKokkos<LMPDeviceType>;

#ifdef KOKKOS_ENABLE_CUDA

#ifdef LMP_KOKKOS_GPU

template class AngleCharmmKokkos<LMPHostType>;

#endif

}

namespace LAMMPS_NS {

template class AngleClass2Kokkos<LMPDeviceType>;

#ifdef KOKKOS_ENABLE_CUDA

#ifdef LMP_KOKKOS_GPU

template class AngleClass2Kokkos<LMPHostType>;

#endif

}