Wesley Maxey

We'll place this on a milestone once we get gpuCI, that's a prerequisite to this as we don't have the capacity on DVS.

Added a test to enforce this and fix in #219...

I don't really have an upshot for you @karthikeyann. A possible solution is to replace: `rmm::device_uvector tile_status;` with: ``` rmm::device_uvector tile_status; using atom_accessor = cuda::atomic_ref; ``` However, it looks this...

I would like to do this after the NVC++ rework. Currently we have the bare minimum exposed for `complex`.

Thanks for the report, I'll try and repro this.

There are some docs in this PR: #257

Possibly addressed with this branch. https://github.com/NVIDIA/libcudacxx/tree/feature/rtti_and_exception_exposure

Moving to backlog as #158 will include this change. @mzient Please let me know if you have any issues with exception/RTTI support on any specific compilers.

TODO: Test ports and heterogeneous tests.

> > Must you force push? As you can see, doing so disassociates the comment history from the code. > > Sorry... I did it only because I rebased on...