Toolsslug: gpu-fractioning-operator

GPU Fractioning Operator

Kubernetes operator enabling GPU memory fractioning with scheduling integration, quotas and metrics.

Curated Intelligence

1 items · Page 1/1
AWS Machine Learning BlogT1·Infra
58 pts
NVIDIA Run: ai Brings GPU Fractioning to Amazon EKS

NVIDIA Run:ai's GPU Fractioning Operator now runs on Amazon EKS, layering Kubernetes-aware scheduling, per-Pod memory contracts and enforcement, and metrics on top of MPS so multiple teams can share a single GPU. Unlike the baseline device plugin that exposes the whole card as one resource and blocks concurrent Pods, it raises utilization for notebooks and small inference services.