Skip to content
Discussion options

You must be logged in to vote

If you already manage the driver + container toolkit on the host (common on a hand-built k3s lab), the standalone k8s-device-plugin is enough and much lighter. The GPU Operator earns its keep when you want the driver, toolkit, DCGM, MIG config and node feature discovery all managed declaratively as a stack — valuable on fleets you do not hand-provision. Lab: device plugin. Fleet/prod: Operator.

Replies: 1 comment

Comment options

iacker
Jul 10, 2026
Maintainer Author

You must be logged in to vote
0 replies
Answer selected by iacker
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Category
Q&A
Labels
None yet
1 participant