Nvidia Launches New Configuration for AI Chip
On Tuesday, Nvidia unveiled a new configuration for its advanced artificial intelligence chips designed to speed generative AI applications.
The new version of the Grace Hopper Superchip is increasing the amount of high-bandwidth memory to enable the powering of larger AI models. They claim that the configuration is optimized to perform AI inference functions that effectively power generative AI applications, including ChatGPT.
Nvidia’s Grace Hopper Superchip design currently stitches together one of the company’s H100 graphics processing units (GPU) with an Nvidia-designed central processor. A larger memory allows the model to remain resident on a single GPU without degrading performance, rather than needing multiple systems or GPUs to run.
The new configuration, GH200, will be available in two versions in the second quarter of next year. One version will include two chips that can be integrated into systems and another as a complete server system combining two Grace Hopper designs.
