At the SIGGRAPH event, Nvidia unveiled its next-generation GH200 Grace Hopper Superchip platform, equipped with the world’s first HBM3e memory, designed for artificial intelligence and high-performance computing. The GH200 Grace Hopper is set to feature the same Grace CPU and GH100 Hopper GPU as Nvidia’s most powerful and popular AI product, the H100, but with three times the bandwidth and over three times the memory capacity.

The ambitious GH200 Grace Hopper Superchip is built upon a 72-core Grace CPU with 480 GB ECC LPDDR5X memory, paired with a GH100 processing GPU boasting six 24 GB stacks and utilizing a 6,144-bit memory interface for a total of 141 GB of HBM3E memory. While Nvidia physically installs 144 GB of memory, only 141 GB is accessible for better efficiency. Since this platform is a dual configuration, these values need to be multiplied by two.

Nvidia’s existing GH200 Grace Hopper Superchip platform comes with 96 GB of HBM3 memory and provides less than 4 TB/s bandwidth. In contrast, the new model increases memory capacity by about 50% and boosts bandwidth by over 25%. These substantial improvements allow the new platform to handle larger AI models and deliver tangible performance enhancements, particularly important for training.

According to Nvidia, the HBM3-equipped GH200 Grace Hopper platform is currently in production and will be commercially available starting next month. On the other hand, the HBM3e-equipped GH200 Grace Hopper platform is currently being sampled and is expected to be released in the second quarter of 2024. Pricing details have not been disclosed, but it’s worth noting that these platforms can be valued in the tens of thousands of dollars depending on the configuration.

Nvidia holds a near monopoly in productive AI-enabled GPUs. Cloud providers like AWS, Azure, and Google all utilize Nvidia’s H100 Tensor Core GPUs. Additionally, Microsoft and Nvidia have formed a partnership to build new supercomputers, although it’s known that Microsoft also aims to develop its own AI chips. Nvidia is also facing competition from AMD, who plans to ramp up production of its own AI GPU in the fourth quarter of this year.

Final Thoughts

Hello #LavegaCrew,

At Lavega Games, we’re passionate about gaming and technology, and we’re excited to share the latest news with you. We understand that a high-quality gaming experience is important, which is why we strive to provide valuable insights and recommendations based on our extensive industry knowledge and connections. Don’t forget to connect with us on our social media channels where you can stay up to date on our latest projects and share your feedback with us. (Instagram, Reddit, Twitter, and GitHub).

Thank you for being a part of our community!

If you are interested in NFTs and Meta world, don’t forget to check out our LavegaPro website.