Posted On: Jul 26, 2023

Today, Amazon AppStream 2.0 announces Graphics G5 instances, which are built on the EC2 G5 family, designed to cater to graphics-intensive applications and machine learning workloads. Amazon EC2 G5 instances feature NVIDIA A10G Tensor Core GPUs and 2nd generation AMD EPYC processors. These instances can be used for graphics-intensive applications, machine learning inference, and deliver higher performance for training simple to moderately complex machine learning models when compared to Amazon EC2 G4dn instances.

AppStream Graphics G5 offers 7 different instance sizes, ranging from xlarge with 4 vCPUs and 16 GB of system memory to 24xlarge with 96 vCPUs and 384 GB of system memory. Most instance sizes provide 24 GB of GPU memory, while G5.12xlarge and G5.24xlarge offer 96 GB of GPU memory.

Graphics G5 instances support Linux and Windows and are available in 9 AWS Regions, including US East (N. Virginia), US West (Oregon), Canada (Central), Europe (Ireland, Frankfurt, London), Asia Pacific (Tokyo, Mumbai), and South America (Sao Paulo). AppStream 2.0 offers pay-as-you-go pricing, see Amazon AppStream 2.0 Pricing for more information.

To get started, select an AppStream 2.0 Graphics G5 instance when launching an image builder or creating a new fleet. You can launch Graphics G5 instances using either the AWS management console or the AWS SDK. To learn more, see AppStream 2.0 Instance Families.