This is The Way- A Crash Course on the Intricacies of Managing CPUs in K8s
summary
Covers CPU management in Kubernetes (kubelet) across setups from single-socket to heterogeneous multi-socket clusters, including proximity-aware allocation for AI/ML workloads near GPUs, core-level pinning inside containers, and reserving cores outside Kubernetes. Explains static and dynamic CPU pinning techniques, practical use cases (AI/ML, telco/5G, networking) and current Kubernetes limitations, and outlines how to surface new use cases to the Kubernetes community.
AI-written summary · about
related talks
23mThis is The Way: A Crash Course on the Intricacies of Managing CPUs...
24mPropagating Programming Paradigms: Lifting High Performance Compute Into K8s
34mUnlocking Performance: Topology-Aware CPU Scheduling W...
13m79 Lightning Talk Dancing with Cores A Path For FineGrained Core Positioning within Kubernetes
30mDeep Dive: CRI- RM Based CPU and NUMA Affinity to Achieve AI Task Accele...
01mKubernetes: What you need to know ... and why
29mKubernetes Workload Resiliency in Action: Beyond Basics
34mWhy and how to build Kubernetes clusters with mixed CPU architectures