Hello Ampere Developers,
Webinar: AI in the Wild: Tracking Bears with Facial Recognition
This month on our Ampere Developer Impact Live webinar, Ed Miller, Senior Principal Engineer at Arm and Technical Lead for the BearID project, joins Ampere’s Chief Evangelist Sean Varley to discuss how the project has enabled tracking the habits of bears in the wild from the Andes to Western Canada. This project uses Ampere CPUs to transform raw wilderness imagery into the actionable data needed to protect biodiversity for future generations.
AI in the Wild- Scaling Bear Facial Recognition with Ampere/Arm64 will be live on July 29th at 9:00 am PDT/noon EDT/5pm BST. Register now to reserve your spot and receive a notification before the event starts - or to watch the event on-demand after it happens.
Jumping on the Arm64 Bandwagon
By now, you may have seen lots of stories of people saving money and getting great performance by using Arm64 instances across the major cloud service providers. In general, Arm64 instances offer the best price-performance across every cloud provider – so why wouldn’t you want to try them out?
However, the idea of moving applications to a new architecture can inspire dread in operations teams. How much effort will application porting require? Will we get the same performance out of the new platform? How much will we have to change our deployment infrastructure?
We have put together a collection of tools and resources to help answer those questions, and others. If you are considering moving workloads from x86 to Arm64, we wanted to share a few practical resources that focus on the actual porting work, what to look for, how to approach it, and tools that can help.
Introducing the Ampere Porting Advisor
The first stop in application migration is evaluating your application to identify potential pitfalls. Are all your dependencies and libraries available on Arm64? Are there known performance issues that can be addressed by moving to more recent application runtimes? Are there parts of the code of your application that will only work on x86?
The Ampere Porting Advisor helps answer these questions and more. Plus, it gives you a checklist of items to be aware of as you start migrating your application. Running the Ampere Porting Advisor against your codebase will give you:
-
Recommendations of application runtimes/toolchain versions
-
Flag any dependencies which are unsupported or have known issues on Arm64
-
Dependencies that may require a native rebuild, but should still work
-
Architecture specific instructions in your code that may need Arm64 equivalents for comparable performance
For an in-depth description of how to get and run the Ampere Porting Advisor, we have an in-depth tutorial here: Ampere Porting Advisor Tutorial
Incremental migration with Kubernetes
If you are running your applications in containers on a hosted Kubernetes service, it is possible to incrementally migrate parts of your application. Simply add some Arm nodes to your cluster, and by using Kubernetes Deployments and workload placement rules, you can have part of your application running on Arm64.
This tutorial walks you through how to migrate a WordPress blog from x86 to Arm64 with zero downtime: How to Migrate to Ampere on OKE with Heterogeneous Kubernetes Clusters
This presentation from ConfigMgmtCamp 2025 by Dave Neary gives the highlights of how to use Argo Rollouts and Argo CD to incrementally migrate applications using Continuous Delivery best practices: Continuous Delivery on multi-architecture Kubernetes clusters with ArgoCD - media.ccc.de
Are you already on the Ampere Developer Community!
Do you have questions about migrating from x86 to Arm64? Join in the conversation on workload migration at the Ampere Developer Community here - share your experiences, or ask any questions you might have.
Join the conversation: https://community.amperecomputing.com
Happy coding and see you in the next edition!
The Ampere Developer Relations Team