GitLab

* Optimized CudaSort for non-uniformly distributed data

* Optimized OpenCLSort for non-uniformly distributed data

* Further tuned distributing elements between buckets

* Copied optimizations over to OpenCL

* Created MonteCarloFlexibleBarostat

* Improved test cases

* Documentation for MonteCarloFlexibleBarostat

* Added a missing include

* Serialization for MonteCarloFlexibleBarostat

* Added MonteCarloFlexibleBarostat to C++ API docs index

* Added citation for MonteCarloFlexibleBarostat

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

* Convert Nose-Hoover into LF middle scheme by copying NH kernels

* Rebrand VelocityVerletIntegrator as NoseHooverIntegrator

* Consolidate NH tests

* NoseHooverChainKernel begone

* Make Windows builds happy

* Add missing header for Windows build

* Fix mistake in CommonKernels header

* Add 6th Yoshida-Suzuki and make it the default

This patch creates a new build in the Travis-CI matrix that compiles and
runs the OpenCL tests using the CPU OpenCL device (AMD drivers). Because
not all of the tests can complete in the allotted 50 minutes, the ctest
invocation is changed to run them in a random order.

This also adds a new executable, TestOpenCLDeviceQuery, that prints
out information about all of the OpenCL devices on the system.