Merging work from ECE1782

Highlights:
Implemented mesh exit and internal surface enter/exit events. Created all the same classes as in SW (e.g. TetraCUDAInternalKernel)
Fixed memory alignment issue with RNG causing TCL exports to fail assertion
Added/Removed some helper functions from CUDAAccel
LOTS of comments/documentation
Exposed m_columns of TextFileMatrixWriter to the user so they can set how many columns of data they want in the output file
Added more stuff to .gitignore