DGX B200 is a fully optimized hardware and software platform that includes the complete NVIDIA AI software stack, including NVIDIA Base Command and NVIDIA AI Enterprise software, a rich ecosystem of third-party support, and access to expert advice from NVIDIA professional services.

The latest addition to the NVIDIA DGX systems, DGX B200 delivers a unified platform for training, fine-tuning and inferencing in a single solution optimized for enterprise AI workloads, powered by the NVIDIA Blackwell GPU.

B200 GPU的晶体管数量是现有H100的两倍多，但B200封装了2080亿个晶体管（而H100/H200上为800亿个）。 这意味着B200芯片封装密度比H100进一步提高，对管理散热和功耗也提出了更高的要求。 引入一种新的计算精度，位宽比FP8进一步降低，B200峰值算力达18P。 位宽介于FP4和FP8之间。 B200有两个Die，高速连接通道 NV-HBI 达到10TB/s。 两个Die是一个统一 …

Mar 18, 2024 · B200 is still about 10% slower than what the fastest Blackwell product can achieve. And what is that peak-performance product? The Grace Blackwell Superchip, GB200.

Feb 19, 2025 · The NVIDIA DGX™ B200 System is the universal system purpose-built for all AI infrastructure and workloads from analytics to training to inference. The system is built on eight NVIDIA B200 Tensor Core GPUs.

Mar 18, 2024 · Nvidia revealed its upcoming Blackwell B200 GPU at GTC 2024, which will power the next generation of AI supercomputers and potentially more than quadruple the performance of its predecessor.

B200 SXM 192 GB combines two graphics processors to increase performance. It features 16896 shading units, 528 texture mapping units, and 24 ROPs, per GPU. Also included are 528 tensor cores which help improve the speed of machine learning applications.

Jun 26, 2024 · See NVIDIA Blackwell GPUs B100 and B200 specifications and comparison to other GPUs used in ML training and inference.

Nov 23, 2024 · NVIDIA B200: The Versatile Performer. The NVIDIA B200 GPU is designed for flexible performance, bridging the gap between high-end training and real-time inference. Key Features of the B200: Optimized Compute Units: Balances high compute power with memory efficiency for mid-scale AI applications.

Mar 19, 2024 · NVIDIA on Monday, at the 2024 GTC conference, unveiled the "Blackwell" B200 and GB200 AI GPUs. These are designed to offer an incredible 5X the AI inferencing performance gain over the current-gen "Hopper" H100, and come with four times the on-package memory.