• TWEAK
• Hardware
• Processor - cpu

First NVIDIA DGX H200 for OpenAI

Translate from : Første NVIDIA DGX H200 til OpenAI

The successor to the highly sought-after H100, the Nvidia DGX H200 boasts 1.4 times more memory bandwidth and 1.8 times more memory capacity. Nvidia CEO Jensen Huang personally delivered the first NVIDIA DGX H200 to OpenAI. In a post shared by Greg Brockman, president of OpenAI, Huang is seen posing with Brockman and boss Sam Altman, with the DGX H200 in the middle.

"The first DGX H200 in the world, delivered to OpenAI and dedicated by Jensen "to the advancement of AI, computing and humanity," Brockman wrote on Twitter. Huang delivering the first GPU to OpenAI seems to have become a new tradition .

Back in 2016, Huang donated the first DGX-1 AI supercomputer to OpenAI, supporting the democratization of AI technology. Back then, it was Elon Musk who received it. The new GPU could be a much-appreciated addition to OpenAI's arsenal, as the organization is currently working on GPT-5 and plans to make Sora publicly available this year.

Nvidia introduced the DGX H200 last year. The upgraded GPU, which succeeds the highly sought-after H100, has 1.4 times more memory bandwidth and 1.8 times more memory capacity. These improvements significantly improve its ability to handle demanding generative AI tasks. In addition, the H200 has a faster memory specification called HBM3e that raises its memory bandwidth to 4.8 terabytes per second from the H100's 3.35 terabytes per second.

Its total memory capacity also increases to 141GB, up from its predecessor's 80GB. " To create intelligence with generative AI and HPC applications, massive amounts of data must be efficiently processed at high speed using large, fast GPU memory," said Ian Buck, vice president of hyperscale and HPC at NVIDIA. "With the NVIDIA H200, the industry's leading AI supercomputing platform has just become faster to solve some of the world's most important challenges."

Nvidia has also launched a new AI supercomputer with the H200. The Nvidia DGX H200 uses NVLink interconnect technology along with the NVLink Switch System, which fuses 256 H200 superchips into a single GPU unit.

This setup achieves an impressive performance of 1 exaflop and offers 144 terabytes of shared memory, marking a significant leap from the previous generation Nvidia DGX A100 introduced in 2020, which had significantly less memory."