(GPU Testing Part 2) Device-compatible Tuple implementation.

An analogue of std::tuple that works on device.

Context: I've tried std::tuple in various versions of NVCC and clang, and although code seems to compile, it often fails to run - generating "illegal memory access" errors, or "illegal instruction" errors. This replacement does work on device.

This is part 2 of a set of changes to simplify creating generic GPU tests.