GitLab

* Vectorize NonbondedForce with no cutoff

* Vectorize CustomNonbondedForce with no cutoff

* Memory efficient dense neighbor list

* Fixed errors

* Began vectorizing CustomNonbondedForce

* Refactored CpuCustomNonbondedForce to support multiple vector sizes

* AVX implementation of CpuCustomNonbondedForce

* Fixed compilation errors

* Reference implementation of computed values for CustomNonbondedForce

* CPU implementation of computed values for CustomNonbondedForce

* Common implementation of computed values for CustomNonbondedForce

* Serialization of computed values

* ForceField supports computed values

The exclusion mask was 8-bits wide, but future CPU support will require
more bits than this (e.g., AVX-512 will be 16-bit).

Created API and reference implementation of restricting CustomNonbondedForce to selected interaction groups

Platform specific header files get installed.  This allows plugins to be built with just an OpenMM installation, not a full source tree.

Redesigned CustomNonbondedForce to eliminate combining rules and allow the expression to directly use parameters for individual particles.