Kubeflow - Works On Arm Sponsored

[Jerry-Ge]

Name, email, company, job title

Jerry Ge @Jerry-Ge, yuzhouge@berkeley.edu (preferred) & jerry.ge@arm.com, UC Berkeley and Arm, Software Engineer Co-applicant: Henry Wang @MrXinWang, henry.wang@arm.com, Arm, Software Engineer

Project Title and description

Kubeflow - Kubeflow is the machine learning toolkit for Kubernetes.

Describe your use case for these machines

• According to the issue - https://github.com/kubeflow/kubeflow/issues/2337. There is a stream of developers are waiting for the enablement of Kubeflow on arm servers.
• I will enable the Kubeflow on arm servers and do benchmarking jobs for some ML workloads

Which members of the community would benefit from your work?

• Kubeflow and Kubernetes users and developers.

Is the code that you’re going to run 100% open source?

Yes. https://github.com/kubeflow/kubeflow

Does this project require ARMv8.2 (Yes/No/Not Sure)? If Yes, please provide details

Not Sure

What infrastructure (computing resources and network access) do you need?

Prefer 2 new Altra Mt. Snow in NA region

Describe / Name the continuous integration (CI) system for this project.

Prow

Please state your contributions to the open source community and any other relevant initiatives.

I am a member of Kubernetes project SIG scheduling. Actively contributing code to scheduler. I would like to make more contributions to the Kubeflow project as well on Arm platforms.

[Jerry-Ge]

/cc @vielmetti

[Jerry-Ge]

/cc @pgmwoa

[pgmwoa]

We are in the process of getting the servers ready for you. You will get an email as soon as the hardware is reserved and ready for use.

[pgmwoa]

Necessary infrastructure is reserved. Please refer to the welcome email for instructions on how to use the reserved server and the supporting resources. We will be looking forward for your feedback / experience on new servers.