The NVIDIA H100 NVL arrives designed specifically for workloads like ChatGPT

Today the NVIDIA GTC took place, a Keynote in which some of the advances that the company and the industry have made both graphically and at the hardware and software level, referring quite a bit to the artificial intelligence that has grabbed so many headlines in recent months. However, as part of thishas been used to launch the NVIDIA H100 NVL.

This particular graphics card comes with a performance 10 times higher than the NVIDIA HGX A100 in terms of text generation speed, something that comes thanks to what NVIDIA mentions as the new Super Chip Grace Hopperthat it has a processor and a GPU on the same PCB, which also houses 94GB of HBM3 memory.

These cards arrive in a format that can be installed in servers that already support graphics cards, so that existing data centers can be upgraded to bring them up to date with the new era of AI text generation, either with ChatGPT or any other artificial intelligence that we may come across in the future.

Additionally, the NVIDIA L4 has been showna card designed specifically for video decoding and encoding, moderation, and video generation using artificial intelligence, as well as real-time video translation and any other use that revolves around the use of artificial intelligence in video.

For her part, the NVIDIA L40 arrives with another purpose in mind, and it is mentioning Omniverse, graphics rendering, and generative artificial intelligences -such as Stable Diffusion- as the workloads for which this GPU will be responsible.

End of Article. Tell us something in the Comments!