Amazon ECS Managed Instances now supports AWS Trainium and AWS Inferentia

Amazon ECS Managed Instances Updates

Amazon ECS Managed Instances now supports AWS Trainium and AWS Inferentia, purpose-built AI accelerators for scalable performance and cost efficiency in generative AI workloads.

ECS Managed Instances is a fully managed compute option that eliminates infrastructure management overhead while providing access to the full capabilities of Amazon EC2. It helps you quickly launch and scale workloads, enhancing performance and reducing total cost of ownership.

New Features

• Support for AWS Trainium and AWS Inferentia in ECS Managed Instances.
• Ability to create an ECS Managed Instances capacity provider and select accelerated instance types.
• NEURON_CORE=all configuration in the ResourceRequirement section of your task definition to allocate all accelerator resources to your workload.

What to do

• Create an ECS Managed Instances capacity provider.
• Select desired accelerated instance types.
• Add NEURON_CORE=all configuration to your task definition.

Source: AWS release notes

If you need further guidance on AWS, our experts are available at AWS@westloop.io. You may also reach us by submitting the Contact Us form.