GitLab

When computing kinetic energy, make sure the shifted velocities are perpendicular to the constraints

Reduced memory use for random numbers, eliminated unnecessary calculation when generating random numbers

Fixes nvcc terminals popping up during CUDA JIT compilation due to invoking system("nvcc ..."). Replaced with MSVC's native CreateProcess() function. 

CudaApplyConstraintsKernel::apply() did not set the context as current, leading to errors when using multiple contexts

No commit message

Fixes uninitialized bits occurring in the PaddedNumAtoms regions by properly constructing them. This should hopefully get rid of some of the NaN errors we've been seeing.

Fixed potential energy discrepancies in large systems between CUDA and Reference platform. The energy calculation now uses the same method as that of reciprocalConvolution so it mirrors the reference implementation. OpenCL is fine.  

When converting to fixed point, multiply by 0x100000000 instead of 0xFFFFFFFF.  This should be (very very slightly) more accurate, since its reciprocal can be exactly represented in floating point.