Eos Supercomputer Fuels Innovation

Providing a glimpse into the infrastructure behind advanced AI factories, NVIDIA has released a video showcasing Eos, its latest data-center-scale supercomputer. Eos, a massive NVIDIA DGX SuperPOD, serves as the platform for NVIDIA developers to create AI breakthroughs using optimized software and accelerated computing infrastructure. With 576 NVIDIA DGX H100 systems and NVIDIA Quantum-2 InfiniBand networking and software, Eos delivers a remarkable 18.4 exaflops of FP8 AI performance. Unveiled at the Supercomputing 2023 trade show, Eos exemplifies NVIDIA's dedication to advancing AI technology.

Optimized for AI Workloads

Each DGX H100 system in Eos is equipped with eight NVIDIA H100 Tensor Core GPUs, resulting in a total of 4,608 H100 GPUs. This enables Eos to handle the most demanding AI workloads, including training large language models, recommender systems, and quantum simulations. Eos serves as a testament to the capabilities of NVIDIA's technologies when operating at scale. Its arrival comes at a crucial time when generative AI is transforming various industries, from drug discovery to chatbots to autonomous machines. To achieve these breakthroughs, organizations require more than just AI expertise and development skills—they need an AI factory. Eos serves as a purpose-built AI engine that is always available and capable of scaling AI model development. Ranked as the 9th fastest supercomputer in the TOP 500 list, Eos pushes the boundaries of AI technology and infrastructure. It incorporates NVIDIA's advanced accelerated computing and networking, as well as sophisticated software offerings like NVIDIA Base Command and NVIDIA AI Enterprise.

Optimized Architecture for AI Workloads

Eos's architecture is specifically designed for AI workloads that demand ultra-low-latency and high-throughput interconnectivity across a large cluster of accelerated computing nodes. This makes it an ideal solution for enterprises seeking to scale their AI capabilities. Powered by NVIDIA Quantum-2 InfiniBand with In-Network Computing technology, Eos's network architecture supports data transfer speeds of up to 400 Gb/s, facilitating the rapid movement of large datasets crucial for training complex AI models.

The Groundbreaking DGX SuperPOD Architecture

At the core of Eos lies the groundbreaking DGX SuperPOD architecture, powered by NVIDIA's DGX H100 systems. This architecture is designed to provide the AI and computing fields with fully integrated full-stack systems capable of operating at an enormous scale. As enterprises and developers worldwide strive to harness the power of AI, Eos stands as a pivotal resource, promising to accelerate the journey towards AI-infused applications that drive every organization forward.