Description
NVIDIA DGX B200 provides a powerful infrastructure for AI applications, offering unparalleled advantages in terms of both performance and sustainability.
Superior Performance
- Advanced GPU Architecture: Equipped with NVIDIA Blackwell Tensor Core GPUs, it delivers 72 petaFLOPS of instruction and 144 petaFLOPS of inference performance.
- NVLink Technology: It enables ultra-high-speed data transfer between GPUs, making it ideal for large language models (LLM) and complex workloads.
2. Efficiency and Sustainability
- Energy Efficiency: It provides up to 50% energy savings with optimum energy consumption. It is also suitable for sustainable AI projects with a maximum power consumption of 14.3 kW.
- Thermal Management: Improved thermal design ensures stable operation even during high performance.
3. Flexible Infrastructure
- Storage and Memory: Up to 4 TB of system memory and 1,440 GB of GPU memory provide ample resources for complex data processing tasks.
- Modular Networking: With InfiniBand and Ethernet support up to 400 Gb/s, it meets high-speed data transfer requirements.
4. Comprehensive Software Ecosystem
- NVIDIA AI Enterprise and Base Command: Offer software tools optimized for all AI workflows. Provides maximum efficiency in all processes, from training to inference.
5. Versatile Applications
- LLM and NLP: Designed for challenging projects such as large language models and natural language processing.
- AI Transformation: Accelerates AI-driven transformations and simplifies processes for organizations.
DGX B200 appeals to a wide range of users, from small businesses to large corporations, with its scalability and flexibility.
Additional information
| Graphics Card | B200 |
|---|---|
| GPU-GPU | SXM |
| SXM Platform | DGX |
Brand
NVIDIA
NVIDIA is a global leader in graphics processing technology, known for its cutting-edge GPUs, AI solutions, and innovations that power gaming, creative work, data centers, and autonomous systems.
