HPE 4-way NVLink Bridge for H200 NVL

Do you need higher performance for training and inference of artificial intelligence (AI), high-performance computing (HPC), or graphics? NVIDIA accelerators for HPE help address the world's most pressing scientific, industrial, and business challenges using AI and HPC. Visualize complex content to develop innovative products, tell compelling stories, and reinvent the cities of the future. Extract new insights from vast amounts of data. Hewlett Packard Enterprise servers with NVIDIA accelerators are designed for the era of elastic computing, offering unparalleled acceleration at any scale.

New features – Additional information: The NVIDIA H200 NVL is suitable for air-cooled rack designs for enterprises that require flexible configurations. With up to four GPUs connected via NVIDIA NVLink and a 1.5x memory expansion, LLM inference can be accelerated by up to 1.7 times and HPC by up to 1.3 times on the H200 NVL compared to the H100 NVL. The NVIDIA H200 NVL includes a five-year NVIDIA AI Enterprise subscription, simplifying the setup of your AI-capable enterprise platform. The 3rd generation NVLink doubles the direct GPU-to-GPU bandwidth. 3rd generation RT cores accelerate rendering workloads. Confidential NVIDIA Computing is an integrated security feature of the Hopper-based accelerators.