GitLab

Export CudaFreeEnergyKernelFactory symbols to avoid link error in test program.

Turn default to ON for nightly building Cuda FreeEnergy plugin, since it is desired for pyopenmm release.

Still doesn't build on windows, but might help with linux (todo next)

Cast exponent in pow() function to RealOpenMM, to avoid overload ambiguity error on Windows when RealOpenMM is defined as double.

Associated code will be removed once confirmed deletions do not break builds

Fixed uninitialized value problem with setting box vectors.  This caused sporadic multithreading issues on Linux platforms.

Use a consistent (and more accurate) value for electric constant everywhere.  Also cleaned up handling of constants and units in various places.

Removed static variables in CudaFreeEnergyKernels.cpp
Moved declaration of global feSim to within method where actually used  in kCalculateNonbondedSoftcore.cu

windows export macro on ReferenceFreeEnergyKernelFactory, to avoid link error on TestReferenceGBVISoftcoreForce

Replace Platform::getPlatform by Platform::getPlatformByName where necessary, due to recently changed interface.

Removed cruft from Cmake file
Modified the way platform is specified in TstFreeEnergyCudaUsingParameterFile.cpp

removed some cruft

Fixed bug: was calling kReduceObcGbsaSoftcoreBornForces instead of kReduceGBVISoftcoreBornForces in GB/VI force calculation

Mods to get Free energy plugin libs to compile on Windows; still problem w/ kCalculateLocalSoftcoreForces.cu
and hence OpenMMFreeEnergyCuda lib

Added #include <typeinfo> to see if it removes errors associated w/ std::bad_cast on Linux Ubuntu 4.3.2

Changed ReferenceFreeEnergyPlatform name to ReferenceFreeEnergy
Removed references to ParameterFileName in GB/VI

Removed references to softCoreLJLambda in NonbondedSoftcoreForce

Removed print cruft

Free energy plugin added
Plugin will not run w/ Obc or GB/VI forces unless line 2004 of gpu.cpp (gpu->sim.totalNonbondOutputBuffers  = 2*gpu->sim.nonbondOutputBuffers;) is commented in -- working on removing this constraint
Also unit tests for GB/VI currently fail