Maintaining OpenCL packages

I tried asking on slack, but since I didn’t get many replies, I am asking here.

Could someone from the JuliaGPU organization please take care of the PRs in OpenCL? There are many attempts to port the package to Julia v1.0 that are hanging there. For example:


If maintainers could share their plans for the projects, I appreciate it. Alternatively, I don’t mind gaining write access to maintain these packages myself. I depend on them for my own work.

Thank you,

5 Likes

Are there any updates on this? Are there any plans to continue to support openCL or alternative packages that I am not aware of?

Yes, the appropriate PRs have been merged for Julia 1.x support, the only issue is fixing precompile (but that’s not strictly necessary, and I don’t know how to make the necessary changes).

Are there any other specific PRs that you want to see merged?

Thank you. No I did not think of anything specific. I noticed the last commit in Nov 2018 and incorrectly assumed the package was not kept up-to-date.

1 Like

No, I think you’re right that it’s not been kept up to date too well. I think CUDAnative and CuArrays stole all the thunder of OpenCL.jl so it hasn’t attracted as many contributors recently. Still, OpenCL support in Julia is important to many of us, so please do contribute PRs or raise issues where necessary. I’d be happy to review and test any PRs you contribute :smiling_face_with_three_hearts:

1 Like

So what’s the situation here then? Is openCL working for 1.x or not? I would like to port my package (for which it is an essential dependency) to 1.x, but am blocked by this non-functionality.

OpenCL.jl should work fine on Julia 1.x. CLArrays is effectively defunct on Julia 1.x, but if you only need support for hand-rolled kernels, then OprnCL.jl should suit your needs.

1 Like