Announcing preview of Amazon EC2 P4de GPU instances for ML training and HPC

We are excited to announce preview of Amazon EC2 P4de instances, our latest GPU-based instances that provide the highest performance for machine learning (ML) training and High-Performance Computing (HPC) applications such as object detection, semantic segmentation, natural language processing, recommender systems, seismic analysis, and computational fluid dynamics.

P4de instances are powered by 8 NVIDIA A100 GPUs with 80GB high-performance HBM2e GPU memory, 2X higher than the GPUs in our current P4d instances. The new P4de instances provide a total of 640GB of GPU memory, which provide up to 60% better ML training performance along with 20% lower cost to train when compared to P4d instances. The improved performance will allow customers to reduce model training times and accelerate time to market. Increased GPU memory on P4de will also benefit workloads that need to train on large datasets of high-resolution data.

P4de instances are available in the AWS US East (N. Virginia) and US West (Oregon) regions. They are available in the p4de.24xl size, providing 96 vCPUs, 8 NVIDIA A100-80GB GPUs, 1.1 TB system memory, 8 TB local NVMe-based SSD storage, 19 Gbps EBS bandwidth, and 400 Gbps networking bandwidth with EFA and GPUDirect RDMA. P4de instances are deployed in EC2 UltraClusters which providing petabit-scale non-blocking networking infrastructure and high-throughput, low-latency storage through FSx for scale-out applications in ML training and HPC. With the new P4de instances we continue to add to the industry’s widest portfolio of accelerated compute instances, featuring platforms powered by our own silicon and by accelerators from our partners, to provide the highest performing NVIDIA GPUs for customers to build, train, and deploy machine learning models at any scale.

To learn more about P4de instances and request access please visit our P4 product page.