GitLab

Separated hardware query functions into a separate file.  CPU platform checks for the correct hardware.

Changed CMake scripts to work correctly if the FFTW threading support was built into the main library instead of a separate library

AmoebaMultipoleForce did not report particle groups, which could potentially lead to incorrect reordering

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

Removed AVX code, since it had very little effect on performance and would have required a more complicated build process.  Also worked around a compilation error with clang.

Converted the array containing atom block indices for the neighbor list from ushort2 to int.  This removes the hard limit of 2 million atoms.

Integrators control when atom reordering gets done, instead of having it happen automatically when forces are evaluated.  This simplifies some code, and also improves unintuitive behavior with CustomIntegrator.