pciolkosz

Results 17 comments of pciolkosz

We should have a `async_buffer` -> `synchronous_buffer` conversion. Should not be any issues assigning the resource this way (`any_async_resource` to `any_resource`)

I agree with the solution. From internal discussion `cuda::device` namespace was added at a time where everything in `cuda` namespace was host/device and we needed a new namespace for device-only...