GPU‑Accelerated Workloads on KubeVirt: Scaling ML/AI in Kuberne...
summary
KubeVirt enables GPU‑backed VMs for ML inference by installing a GPU device plugin such as NVIDIA device-plugin and GPU operator, ensuring host drivers and CUDA are present, then declaring GPU resource requests in a VM YAML so Kubernetes schedules a node and does device pass-through to guest OS where the guest sees physical GPUs for training or inference. It flags live migration limits and setup complexity and hooks into Prometheus/Grafana monitoring for GPU metrics.
AI-written summary · about
chapters (6)
- 0:00GPU workloads intro
- 1:50What is KubeVirt
- 3:20Why GPUs for AI/ML
- 6:09Enabling GPUs in K8s
- 8:17GPU config steps
- 10:21Key takeaways
related talks
31mKubernetes For GPU Powered Machine Learning Workloads In Academia
31mBut we have GPUs at home - on-prem ai with kubernetes, Openstack & KubeVirt
44mGPU Sharing for Machine Learning Workload on Kubernetes
06mKubernetes Love Machine Learning, Even on Private Cloud
30mRunning AI Workloads in Containers and Kubernetes
33mAI workloads on Kubernetes - How to maximize GPU utilization and cut costs
27mHarnessing Kubernetes for Scalable AI/ML Workloads
32mScaling AI Workloads with Kubernetes: Sharing GPU Resources Across Multiple Containers