Skip to content

Conversation

@RainerHeintzmann
Copy link
Contributor

This second attempt, uses a very light-weight implementation to add CuArray support to ShiftedArrays.jl.
It is based on using an Extension Package, such that ShiftedArrays does not drag in any extra packages, but if CUDA.jl is present, the adaptation is used.
Note that the show() function was specified by CuArrays preventing errors via the @allowscalar macro.
The tests were extended to CuArray usage, but this has to be enabled manually in the runtests.jl file.

@piever
Copy link
Collaborator

piever commented Dec 20, 2023

Thanks for the PR, I've left a review with some suggestions. I'll be traveling for the holidays so I might be slow in responding for the next couple of weeks.

@RainerHeintzmann
Copy link
Contributor Author

... there were a couple of typos. Now the show adaptation is commented out and the typos were fixed. The tests run fine. Seems we are getting closer.

@RainerHeintzmann
Copy link
Contributor Author

RainerHeintzmann commented Feb 12, 2025

Can we move ahead as is (with the optional CUDA testing)? If not, we can also make the tests CUDA-dependent.
I updated the CUDA support now to a recent version (5.1 upwards to 5.6), since CUDA.jl changed the broadcast type.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants