Jan Stephan

We will try and tackle this for 0.9 since it was mentioned by CMS as nice-to-have.

We have a new dependency: Boost.Atomic. Right now this is non-optional but can easily be turned into an optional dependency.

@psychocoderHPC Is this still something you want to pursue?

Our internal discussion didn't show much love for the fiber back-end. I'll submit a PR to remove the back-end. Until then this issue is open for anyone who wants to...

> There once was a ticket to scale this across multiple threads, not sure where this went. This should be #22.

While we are talking about legality: Is it even legal (from OpenMP's point of view) to have an activated OpenMP back-end together with the CUDA back-end which might spawn a...

Maybe let the default constructor construct an empty buffer? In this case we should add an `empty()` check to the buffer interface.

If we're approaching this subject anyway we should also discuss the general need for n-dimensional buffers in alpaka. If we ditch pitched allocations, why not ditch 2D / 3D allocations...

One could argue that we should also remove the type from the buffer so that the user always receives a range of `std::byte`s and has to set up the actual...

> Is it allowed for the host code to be multithreaded itself? The SYCL back-end currently assumes this and I think it would be useful to have this guaranteed for...