SageMaker Notebook Instances now support P5en.48xl instance types
Amazon Web Services has announced the general availability of EC2 P5en.48xl instances for SageMaker notebook instances, bringing enhanced GPU capabilities to machine learning development environments. The P5en instances feature eight NVIDIA H200 GPUs, which deliver 1.7x more GPU memory and 1.4x greater memory bandwidth compared to the H100 GPUs found in standard P5 instances. These GPUs are paired with custom 4th Generation Intel Xeon Scalable processors and Gen5 PCIe connectivity that provides up to 4x the bandwidth between CPU and GPU. The new instances also incorporate up to 3200 Gbps of third-generation Elastic Fabric Adapter (EFA) using Nitro v5, delivering up to 35% improvement in latency compared to previous-generation P5 instances. This enhanced networking performance is designed to improve collective communications for distributed training workloads including deep learning, generative AI, real-time data processing, and high-performance computing applications. The P5en.48xl instances are currently available in AWS US East (N. Virginia and Ohio), US West (Oregon), and Asia Pacific (Tokyo) regions.
Why It Matters
This infrastructure upgrade significantly enhances AWS's competitive position in the AI/ML cloud computing market by offering developers access to NVIDIA's latest H200 GPUs through familiar notebook environments. The improved GPU memory capacity and bandwidth, combined with enhanced CPU-GPU connectivity, addresses key bottlenecks in training large language models and other memory-intensive AI workloads. The 35% latency improvement in distributed training scenarios could substantially reduce training times and costs for enterprises developing complex AI models.
This summary is generated using AI analysis of the original press release. Always refer to the original source for complete details.