-
Notifications
You must be signed in to change notification settings - Fork 231
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[FEA] Support RAPIDS Spark plugin on ARM #6881
Labels
feature request
New feature or request
Comments
sameerz
added
feature request
New feature or request
? - Needs Triage
Need team to review and classify
labels
Oct 21, 2022
Let's keep this open We will reapply CICD update after developers refactor modules structures for different cpu archs in plugin and JNI repos. |
UCX tracking issue: openucx/ucx#8762 |
Just to be clear @pxLi:
|
|
merged #9888, udpating internal pipelines. I will update this ticket no other issue |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Is your feature request related to a problem? Please describe.
A clear and concise description of what the problem is. Ex. I wish the RAPIDS Accelerator for Apache Spark would run on ARM
Describe the solution you'd like
Distribute a plugin jar that is functional on ARM-based architecture (including cudf binaries on ARM)
More details:
We need to distribute the spark-rapids-jni jar with 2 different classifers:
And Spark-rapids may have 2 profiles,
one is for amd64 to depend on jni's cuda11-amd64 classifier,
another is for aarch64(arm) to depend on jni's cuda11-aarch64 classifier.
For amd64, we build on centos7.
For Arm, we'll build on ubuntu 18.04.
It would be nice if we can build jni jar with Dockerfile.
https://hub.docker.com/layers/nvidia/cuda/11.5.2-runtime-ubuntu18.04/images/sha256-05d99b36cccb63102dbb1f5d9e74daf07d49a62151175693a6146c6d0244dc10?context=explore
Describe alternatives you've considered
Additional context
The text was updated successfully, but these errors were encountered: