Keynote: High Performance Computing on Google Kubernetes Engine
Talkat Kubernetes Batch + HPC Day Europe 2022 · on Tuesday, 2022-05-17· at 13:15· Room 4F | Event Center·
summary
Google claims GKE runs production HPC at scale, supporting clusters up to ~45,000 nodes and shipping a custom autoscaler that optimizes for customer cost. The platform integrates hardware (GPUs, new CPU architectures), offers GPU slicing and spot VM orchestration, and improves cluster lifecycle with easier creation and upgrades. Roadmap mixes open-source contributions to job/queue operators with infrastructure features; several capabilities are already live and geared toward cost and performance for scientific workloads.
AI-written summary · about
chapters (4)
- 0:00Start
- 2:38How Kubernetes helps?
- 3:24Google Kubemetes Engine
- 5:21Google Kubernetes Engine highlights
related talks
06mKeynote: Open Innovation: Kubernetes on Google Cloud
09mKeynote: Kubernetes for GenAI: The New Operating System of the AI-First World
16mRunning Services at Google Scale
36mProduction GPU Cluster with K8s for AI and DL Workloads
31mManaged Kubernetes — Next Gen Academic Infrastructure?
19mKeynote: Building Bridges: Cloud Native and High Performance Computing
22mKeynote: Kubernetes Project Update
20mKeynote: Kubernetes at GitHub