GitLab

* Use Intel OpenCL for CI

* Set environment variables

* Try to get CI to run

* Debugging

* Debugging

* Fixes for Intel OpenCL

* Unified lots of parallel computation code between platforms

* Unified test code between platforms

* Eliminated duplicated timing code

* Optimize CustomNonbondedForce.updateParametersInContext()

* Optimized uploading changed values to GPU

* Optimized updateParametersInContext() for lots of bonded forces

* Optimized updateParametersInContext() for CustomExternalForce

* Optimized updateParametersInContext() for NonbondedForce

* Code changes for HIP platform

4*(1<<30) evaluates to 0 so the check returns true on 2GB cards. Test runs fine on 2GB cards but requires over 4GB of system memory.

* Fixed range overflow with very large numbers of atoms

* More fixes to overflow with large numbers of atoms

* Fix test failures

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

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.

OpenCLArray is no longer templatized and doesn't provide a host buffer.  This is in preparation for adding mixed/double precision support.

On Fermi, store forces as 64 bit fixed point and update with atomic operations.  Also various other minor optimizations.

Moved SFMT into libraries folder, and restructured it to allow multiple independent random number generators

All platforms now handle nonbonded exceptions in a consistent way (they ignore cutoffs and periodic boundary conditions).  Also started OpenCL implementation of CustomHbondForce.