NVIDIA/k8s-device-plugin

Health v2 · maintenance only

Adoption not part of the score

3859 stars · 853 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

NVIDIA's official Kubernetes device plugin that exposes NVIDIA GPUs as schedulable resources in Kubernetes clusters. It supports GPU sharing via time-slicing and MPS, per-node configuration through ConfigMaps, and deployment via Helm.

Use cases

• schedule GPU workloads in Kubernetes
• expose NVIDIA GPUs as Kubernetes resources
• share GPUs across pods with time-slicing
• run CUDA jobs on a Kubernetes cluster
• set up GPU nodes for machine learning training
• configure per-node GPU settings via node labels

When to choose

• you run NVIDIA GPUs on Kubernetes and need pods to request GPU resources
• you want GPU time-slicing or MPS-based sharing across pods
• you deploy GPU workloads via Helm and want official NVIDIA support

When to avoid

• your cluster has no NVIDIA GPUs
• you use a non-Kubernetes orchestrator like Docker Swarm or Nomad
• you need AMD or Intel GPU scheduling

Facets

plugin · maturity stable

container-orchestration gpu-computing deployment configuration-management cloud-computing gpu-computing machine-learning go nvidia gpu-scheduling device-plugin helm-chart cuda time-slicing mps containers devops kubernetes docker linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="NVIDIA/k8s-device-plugin")