Tag: #MIG
Kubernetes, DevOps and MLOps, written from things we ran and measured ourselves · 1 posts
Multi-tenant GPU Scheduling and Capacity Planning
Starting from the limitations of how the default scheduler handles GPUs, this post outlines the design differences between Kueue and Volcano, topology-aware placement, the isolation limits of GPU sharing, and the current
2026-10-08 · 37 min read #Kubernetes#GPU#Kueue#Volcano#스케줄링