NVIDIA Data Center GPU
NVIDIA Data Center GPUs accelerate demanding HPC and other workloads, enabling data scientists and researchers to analyze petabytes of data far faster than traditional CPUs across a wide range of applications, from energy exploration to deep learning.
NVIDIA Data Center GPU Product List
NVIDIA H200 Tensor Core GPU
The NVIDIA H200 Tensor Core GPUs power generative AI and HPC workloads with market-changing performance and memory capabilities. As the first GPU to feature HBM3e, the H200's large-capacity, high-speed memory accelerates generative AI and large-scale language models (LLMs) while driving scientific computing for HPC workloads.
NVIDIA RTX PRO™ 6000 Blackwell Server Edition
Powered by the innovativeNVIDIA Blackwell architecture, the NVIDIA RTX PRO™ 6000 Blackwell Server Edition integrates cutting-edge AI processing power with exceptional visual computing performance to dramatically accelerate complex workloads in enterprise data centers. Featuring 96GB of ultra-fast GDDR7 memory, the NVIDIA RTX PRO 6000 Blackwell delivers unparalleled performance and flexibility for a wide range of use cases, from agent AI, physical AI, and scientific computing to rendering, 3D graphics, and video processing.
NVIDIA L40
Based on the Ada Lovelace architecture, the NVIDIA L40 delivers innovative neural graphics, visualization, computing, and AI capabilities for GPU-driven data center workloads.
Fan/cooling function
NVIDIA Data Center GPUs use passive cooling and do not have fans, requiring the server itself to have a cooling system. By not having fans, Data Center GPUs reduce the risk of failure, making them ideal for use in data centers and other environments where replacing GPU cards is difficult.
Memory size and memory bandwidth
In the AI training phase, which involves handling large amounts of data, the amount of data transferred to GPU memory is substantial, making the corresponding memory size and memory bandwidth extremely important factors. NVIDIA Data Center GP offers large memory size and memory bandwidth, making it an ideal product for AI/HPC. A table comparing each NVIDIA Data Center GPU is shown below.
|
NVIDIA H200 (PCIe) |
NVIDIA RTX PRO™ 6000 Blackwell Server Edition |
NVIDIA L40 |
|
| Memory Size |
141GB |
96GB GDDR7 |
48GB GDDR6 |
| Memory BW(GB/s) |
4800 |
1597 |
864 |
| Interconnect |
PCIe Gen5 |
PCIe Gen5 |
PCIe Gen5 |