NVIDIA AI Infrastructure and Operations Operations and Monitoring NCA-AIIO
Quiz 9 Question 12 of 20

You are administering a Kubernetes-based AI cluster where researchers run a mix of workloads: some are long-running, multi-GPU training jobs, while others are quick, single-GPU inference tests or high-memory preprocessing steps. To maximize cluster efficiency and keep queue times short, you need to automate resource allocation so that compute and memory resources are assigned based on real-time demands and current node availability. Which of the following scheduling methods is the most appropriate?

Select an answer to reveal the explanation.