AWS Publishes NVIDIA Run:ai GPU Fractioning Guide for Amazon EKS
Start here
AWS's Machine Learning Blog published a walkthrough for configuring NVIDIA Run:ai GPU Fractioning on Amazon EKS, letting multiple Pods share a single GPU under enforced per-Pod memory limits. In the demonstrated setup, a single g5.xlarge instance with an A10G GPU runs two Pods, each with an 8 GiB contract and 6 GiB allocated, while a 9 GiB request is rejected. The setup uses KAI Scheduler and the GPU Fractioning Operator from the open-sourced NVIDIA Run:ai, and requires building a custom Amazon Linux 2023 image. The walkthrough validates only the G5/A10G configuration and requires pre-provisioned GPU capacity. The custom image cannot run on Amazon EKS Auto Mode-managed nodes, because Auto Mode selects and manages their AMI. Following the walkthrough creates billable resources, including an Amazon EKS control plane, a NAT gateway, instances, volumes, and an AMI snapshot.
Generated from reports · updated 2 h ago
Report timeline
Follow the reports to see the event from different sides.
- AWS Machine Learning BlogUsing NVIDIA Run:ai GPU Fractioning on Amazon EKS
This walkthrough shows how to configure NVIDIA Run:ai GPU Fractioning on Amazon EKS so multiple Pods share one GPU with enforced per-Pod memory limits. On a single g5.xlarge with an A10G, two Pods each get an 8 GiB contract and allocate 6 GiB, while a 9 GiB request is rejected. It uses KAI Scheduler and the GPU Fractioning Operator from the open-sourced NVIDIA Run:ai.
Heat for this event
- Current, comparable range
- 9
- Peak, comparable range
- 10November 20 01:00
- Change in 24 hours
- –
The trend compares the same participants over hours observed continuously, so its range can be narrower than the current heat count. Hover or tap the chart for hourly heat; the arrow keys move through it.