C++ Performance Portability Programming

 This implements a programming model in C++ for writing performance
 portable applications targeting all major HPC platforms. For that purpose it
 provides abstractions for both parallel execution of code and data management.
 Kokkos is designed to target complex node architectures with N-level memory
 hierarchies and multiple types of execution resources. It currently can use
 CUDA, HPX, OpenMP and Pthreads as backend programming models with several other
 backends in development.