Announcing Region Expansion of P4de instances on SageMaker Studio notebooks
Amazon Web Services has expanded availability of its high-performance P4de instances to three additional regions for SageMaker Studio notebooks, including Asia Pacific (Tokyo and Singapore) and Europe (Frankfurt). The P4de instances are equipped with eight NVIDIA A100 GPUs featuring 80GB of high-bandwidth memory (HBM2e) each, delivering double the GPU memory capacity compared to the existing P4d instances and providing a total of 640GB of GPU memory per instance. The enhanced instances deliver significant performance improvements for machine learning workloads, offering up to 60% better ML training performance while reducing training costs by 20% compared to P4d instances. The increased GPU memory capacity is particularly beneficial for organizations working with large datasets and high-resolution data, enabling faster model training times and accelerated development cycles. The instances are now accessible through SageMaker Studio's JupyterLab and CodeEditor applications in the newly supported regions.
Why It Matters
This regional expansion addresses the growing global demand for high-performance GPU compute resources needed for large-scale AI and machine learning workloads. The significant performance improvements and cost reductions make advanced AI development more accessible to organizations in key markets, while the doubled GPU memory capacity enables training of increasingly complex models that were previously constrained by memory limitations.
This summary is generated using AI analysis of the original press release. Always refer to the original source for complete details.