NVIDIA H200 GPU

From GM-RKB
Jump to navigation Jump to search

An NVIDIA H200 GPU is a high-end GPU.



References

2023

  • GBard
    • The NVIDIA H200 is a high-end GPU designed for training and deploying artificial intelligence (AI) models. It is based on the NVIDIA Hopper architecture and is the first GPU to offer 141 gigabytes (GB) of HBM3e memory at 4.8 terabytes per second (TB/s) —that's nearly double the capacity of the NVIDIA H100 Tensor Core GPU with 1.4X more memory bandwidth. [Image of H-200 GPU]
    • The H200 also features a new FP8 Tensor Core that delivers 32 petaflops of FP8 performance, which is 2.3X faster than the H100. This makes the H200 ideal for training large language models and other generative AI applications.
    • In addition to its raw performance, the H200 also offers a number of features that make it easier to use and deploy AI models. These features include:
      • A new software development kit (SDK) that makes it easier to develop and deploy AI applications on the H200.
      • A new NGC container registry that provides pre-built and optimized AI models for the H200.
      • A new NVIDIA AI Enterprise software suite that provides tools for managing and deploying AI models in production environments.