[FEA]: Hoist NVRTCC to CCCL level

### Is this a duplicate?

- [ ] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cccl/issues) for this request and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.md)

### Area

General CCCL

### Is your feature request related to a problem? Please describe.

Currently, CUB has to [reimplement](https://github.com/NVIDIA/cccl/blob/main/cub/test/catch2_test_nvrtc.cu) some of the nvrtc testing. Passing include paths from CMake to the nvrtc test can hardly be called reliable. With introduction of cuda next, the duplication of nvrtc hacks will only increase.

### Describe the solution you'd like

I'd like [NVRTCC](https://github.com/NVIDIA/cccl/issues/465) to be accessible and used at CCCL level: libcu++, CUB, cuda next.

### Describe alternatives you've considered

_No response_

### Additional context

_No response_

Start by reading cub/test/catch2_test_nvrtc.cu and the CMake handling for include paths, then review issue #465 for the NVRTCC context. Done means NVRTCC is accessible at the CCCL level and can be used by libcu++, CUB, and cuda next without each component reimplementing NVRTC testing hacks.

Written by the indexing model from the issue text.