Nvidia CPU designed for scalable solutions with PCIe connectivity 144 cores and 4 GPUs for AI acceleration 1.3 TB memory makes NVL4 ideal for data-intensive applications
Nvidia announced the GB200 NVL4, an advanced platform designed to meet the needs of modern data centers and compute workloads.
The company’s decision to introduce the GB200 NVL4 comes after NVIDIA recently restructured its product lineup, setting aside other NVL platforms in favor of new options like NVL4. is.
The GB200 NVL4 combines two Arm-based Grace CPUs and four Blackwell GPUs. It is part of Nvidia’s Grace Blackwell lineup, which aims to provide flexible and scalable solutions for a variety of deployment scenarios.
fast and scalable
The GB200 NVL4 is designed as a mid-range option in the Grace Blackwell family, sitting below high-performance configurations such as the NVL72 and NVL36. Each unit has two Grace CPUs, each with 72 Arm Neoverse V2 cores for a total of 144 cores. Adding four Blackwell GPUs ensures robust acceleration for AI, HPC, and other compute-intensive tasks.
Six MCIO connectors under each CPU provide PCIe connectivity and enable high-speed data transfer. It also supports the integration of NICs, SSDs, and other required components while maintaining efficiency.
Power consumption is a critical factor in modern data centers, and Nvidia estimates that the GB200 NVL4 will consume just over 6 kW of power per server when fully configured. Although this represents a significant power consumption, it is a notable improvement over larger platforms such as the Nvidia DGX-1 and HGX-1, where previous systems consumed approximately 3.5 kW.
Additionally, the device also supports up to 1.3 TB of combined memory, making it a strong candidate for memory-intensive applications while ensuring efficient data handling and processing. NVL4 is expected to fill a niche for organizations seeking a powerful yet relatively low-power solution compared to GPU-rich NVL platforms.
Although the GB200 NVL4 and the higher-end NVL72 look similar at first glance, there are notable differences. NVL72 features edge connectors for spine configurations and is focused on large-scale deployments with twice the computing resources. In contrast, NVL4 is designed as a more compact, power-conscious node while offering superior performance for its size.
