New Amazon EC2 P6e-GB200 UltraServers accelerated by NVIDIA Grace Blackwell GPUs for the best AI efficiency

At this time, we’re saying the final availability of Amazon Elastic Compute Cloud (Amazon EC2) P6e-GB200 UltraServers, accelerated by NVIDIA GB200 NVL72 to supply the best GPU efficiency for AI coaching and inference. Amazon EC2 UltraServers join a number of EC2 cases utilizing a devoted, high-bandwidth, and low-latency accelerator interconnect throughout these cases.

The NVIDIA Grace Blackwell Superchips join two high-performance NVIDIA Blackwell tensor core GPUs and an NVIDIA Grace CPU primarily based on Arm structure utilizing the NVIDIA NVLink-C2C interconnect. Every Grace Blackwell Superchip delivers 10 petaflops of FP8 compute (with out sparsity) and as much as 372 GB HBM3e reminiscence. With the superchip structure, GPU and CPU are colocated inside one compute module, rising bandwidth between GPU and CPU considerably in comparison with present era EC2 P5en cases.

With EC2 P6e-GB200 UltraServers, you possibly can entry as much as 72 NVIDIA Blackwell GPUs inside one NVLink area to make use of 360 petaflops of FP8 compute (with out sparsity) and 13.4 TB of complete excessive bandwidth reminiscence (HBM3e). Powered by the AWS Nitro System, P6e-GB200 UltraServers are deployed in EC2 UltraClusters to securely and reliably scale to tens of hundreds of GPUs.

EC2 P6e-GB200 UltraServers ship as much as 28.8 Tbps of complete Elastic Material Adapter (EFAv4) networking. EFA can be coupled with NVIDIA GPUDirect RDMA to allow low-latency GPU-to-GPU communication between servers with working system bypass.

EC2 P6e-GB200 UltraServers specs
EC2 P6e-GB200 UltraServers can be found in sizes starting from 36 to 72 GPUs underneath NVLink. Listed here are the specs for EC2 P6e-GB200 UltraServers:

P6e-GB200 UltraServers are perfect for essentially the most compute and reminiscence intensive AI workloads, equivalent to coaching and inference of frontier fashions, together with combination of specialists fashions and reasoning fashions, on the trillion-parameter scale.

You possibly can construct agentic and generative AI purposes, together with query answering, code era, video and picture era, speech recognition, and extra.

P6e-GB200 UltraServers in motion
You should use EC2 P6e-GB200 UltraServers within the Dallas Native Zone by way of EC2 Capability Blocks for ML. The Dallas Native Zone (us-east-1-dfw-2a) is an extension of the US East (N. Virginia) Area.

To order your EC2 Capability Blocks, select Capability Reservations on the Amazon EC2 console. You possibly can choose Buy Capability Blocks for ML after which select your complete capability and specify how lengthy you want the EC2 Capability Block for u-p6e-gb200x36 or u-p6e-gb200x72 UltraServers.

As soon as Capability Block is efficiently scheduled, it’s charged up entrance and its worth doesn’t change after buy. The fee might be billed to your account inside 12 hours after you buy the EC2 Capability Blocks. To be taught extra, go to Capability Blocks for ML within the Amazon EC2 Person Information.

To run cases inside your bought Capability Block, you should use AWS Administration Console, AWS Command Line Interface (AWS CLI) or AWS SDKs. On the software program facet, you can begin with the AWS Deep Studying AMIs. These photographs are preconfigured with the frameworks and instruments that you simply most likely already know and use: PyTorch, JAX, and much more.

You can too combine EC2 P6e-GB200 UltraServers seamlessly with varied AWS managed companies. For instance:

• Amazon SageMaker Hyperpod gives managed, resilient infrastructure that mechanically handles the provisioning and administration of P6e-GB200 UltraServers, changing defective cases with preconfigured spare capability throughout the similar NVLink area to take care of efficiency.
• Amazon Elastic Kubernetes Providers (Amazon EKS) permits one managed node group to span throughout a number of P6e-GB200 UltraServers as nodes, automating their provisioning and lifecycle administration inside Kubernetes clusters. You should use EKS topology-aware routing for P6e-GB200 UltraServers, enabling optimum placement of tightly coupled parts of distributed workloads inside a single UltraServer’s NVLink-connected cases.
• Amazon FSx for Lustre file programs present knowledge entry for P6e-GB200 UltraServers on the tons of of GB/s of throughput and hundreds of thousands of enter/output operations per second (IOPS) required for large-scale HPC and AI workloads. For quick entry to massive datasets, you should use as much as 405 TB of native NVMe SSD storage or just about limitless cost-effective storage with Amazon Easy Storage Service (Amazon S3).

Now obtainable
Amazon EC2 P6e-GB200 UltraServers can be found right this moment within the Dallas Native Zone (us-east-1-dfw-2a) by way of EC2 Capability Blocks for ML. For extra info, go to the Amazon EC2 pricing web page.

Give Amazon EC2 P6e-GB200 UltraServers a attempt within the Amazon EC2 console. To be taught extra, go to the Amazon EC2 P6e cases web page and ship suggestions to AWS re:Put up for EC2 or by way of your typical AWS Help contacts.

— Channy