Deploying GPU-Enabled Workloads on Red Hat Enterprise Linux for Workstation on AWS
What if you need a high-end GPU for your workloads but don’t want to buy the hardware, or you need that compute power to travel with you? In this video I walk through deploying Red Hat Enterprise Linux as a GPU-enabled cloud workstation on AWS.
A few things worth carrying away from this one:
- RHEL is the VFX platform of choice. I point out that RHEL 9 was named part of the VFX Reference Platform in 2022, so this isn’t a novelty, it’s the foundation for animation, scientific calculations, and CAD-style demanding work.
- The AMI catalog is the easy path. I use the AWS AMI catalog and search the marketplace AMIs for the RHEL grid image, launch an EC2 instance with mostly defaults, and pick between the grid or Tesla NVIDIA driver sets that Red Hat supports on AWS.
- NICE DCV handles the remote connection. RHEL for Workstation on AWS uses NICE DCV for remote connectivity to those high-end NVIDIA GPUs, and I’m honest about a bit of production trickery: a few command-line steps live in the companion blog post rather than on screen.
Take a world-class distro, a ten-year life cycle, and a cloud GPU, and you’ve got a workstation you can reach from anywhere.
You might also like
Inventory and Identity | Modernizing RHEL Management 09
A close look at the inventory that spans more than just Insights systems, plus how role-based access control shapes what each user can see and do.
Build and Launch | Modernizing RHEL Management 02
An all-demo episode where John builds a custom RHEL image and launches two systems straight into AWS with auto-registration turned on.
Management Techniques 101 | Modernizing RHEL Management 01
The kickoff episode where John Spinks and I set the baseline on what Red Hat Insights is and how to get your systems connected.