Michael Schellenberger Costa

Replace usages of `thrust::optional` with `std::optional`

6

We want to get rid of thrust types in API boundaries so replace them by the better suited std types

[PoC]: Implement `cuda::experimental::vector`

16

closes: https://github.com/NVIDIA/cccl/issues/2057 This is a PoC for a potential `cuda::experimental::vector` Nothing here is production ready. For example all algorithms used currently are serial. The question is what is the exact...

Asynchronous operations are notoriously difficult. We need to ensure that any memory allocation that we got from `cudaMallocAsync` is actually valid when we are trying to access it. Furthermore, we...

[PoC]: Implement `cuda::experimental::uninitialized_async_buffer`

8

The `uninitialized_async_buffer` provides a stream ordered allocation of N elements of type T utilizing a `cuda::mr::async_resource` to allocate the storage. The buffer takes care of alignment and deallocation of the...

Add nvhpc devcontainer to CI

13

We also support the latest nvc++ as a host compiler, so we should test it in CI

Implement P0843 `inplace_vector`

11

This implements `inplace_vector` a resizable container with a fixed capacity that stores its elements in a local array. Due to the fact that everything is local we are more or...

Test and fix failing nightly libcudacxx + CUB jobs

4

Some of the libcudacxx nightly tests are not running because they would fail. We need to fix that and ensure them running fine

Stream ordered memory allocations a incredibly important for SOL applications. At the same time they are easy to misuse. We need to provide uses with higher level data structures that...