[DOC]: Add how-to guide for invoking CUB algorithms from CUDA kernels

### Is this a duplicate?

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

### Is this for new documentation, or an update to existing docs?

New

### Describe the incorrect/future/missing documentation

CUB supports use cases where selected device-wide algorithms are launched from within a CUDA kernel via CUDA Dynamic Parallelism, but the documentation does not currently provide a clear how-to guide or complete example for this workflow.

We should add a CUB documentation page that explains how to invoke a CUB algorithm from device code using CDP, including the required compilation/linking setup, temporary storage handling etc.

Apart from that, we should modify Doxygen macro `cdp_class{1}` to link to this guide.

### If this is a correction, please provide a link to the incorrect documentation. If this is a new documentation request, please link to where you have looked.

_No response_