Announcing Capacity Blocks support for AWS Parallel Computing Service

AWS has announced Capacity Blocks support for the AWS Parallel Computing Service (AWS PCS), allowing organizations to reserve and schedule GPU-accelerated Amazon EC2 instances for future use.

• Customers can reserve 1-64 GPU instances up to 8 weeks in advance
• Supports NVIDIA Hopper GPU series and AWS Trainium instances
• Offers flexible scheduling for urgent and planned GPU workloads
• Enables resource sharing across multiple PCS clusters
• Provides seamless integration with existing HPC and ML workflows

Key benefits include reserved GPU access, flexible scheduling, resource sharing, and easy integration with PCS Compute Node Groups. The feature is ideal for AI, machine learning, and GPU-accelerated scientific computing workloads.