-
Notifications
You must be signed in to change notification settings - Fork 2
Update to Julia@v1.11 & CUDA.jl@v5.5.2 #82
Conversation
maleadt
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think you need more changes than this, e.g., updating Julia in the Dockerfile to match the manifest, updating the preferences (version=local changed to local=true + version=...), etc.
I'm very sure about this point. The Julia version is specified through docker build arguments. To make the Feel free to suggest more improvements. |
| @@ -1,2 +1,3 @@ | |||
| [CUDA_Runtime_jll] | |||
| version = "local" | |||
| __clear__ = ["version"] | |||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why the clear?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I guess this is auto generated?
TBH, I'm not sure how to set the version here.
|
Alternative, we could get rid of the Manifest altogether and simply use compat bounds. That would make the arg the single point of truth. I can take a look if you want (or feel free to experiment yourself). |
|
I've opted to fold these improvements into the existing Dockerfile of the CUDA.jl repository: JuliaGPU/CUDA.jl@6f9bcaa You can now pull the |
I also added a github action to upload the generated docker images. Feel free to adjust it as necessary.