Nvidia reveals liquid-cooled A100 and Grace CPU-powered systems at Computex 2022

Last night during Nvidia's Computex 2022 keynote, the company had a few major announcements to make. On the gaming side, we've already seen the world's first 500Hz monitor with G-Sync and Reflex technology. On the hardware side, Nvidia primarily focused on data centre and accelerators, including a new liquid-cooled A100 GPU, and an update on the Nvidia Grace CPU project. 

The Nvidia A100 graphics card is the company's flagship ‘Ampere' GPU. These have primarily been available with a passive cooler but now, Nvidia is making a liquid-cooled version available for the first time. This will help data centres reduce power costs due to more efficient cooling.

Nvidia is also making its big push into the CPU space with the Nvidia Grace Superchip and Grace Hopper Superchip. Nvidia has four server models lined up, aimed at cloud computing, Omniverse, HPC and AI training. Different Grace servers will be produced by the likes of ASUS, Gigabyte, Supermicro, Wiwynn, QCT and Foxconn, and the first systems will begin shipping in the first half of 2023. Here's a breakdown of each server design:

• NVIDIA HGX Grace Hopper systems for AI training, inference and HPC are available with the Grace Hopper Superchip and NVIDIA BlueField-3 DPUs.
• NVIDIA HGX Grace systems for HPC and supercomputing feature the CPU-only design with Grace CPU Superchip and BlueField-3.
• NVIDIA OVX systems for digital twins and collaboration workloads feature the Grace CPU Superchip, BlueField-3 and NVIDIA GPUs.
• NVIDIA CGX systems for cloud graphics and gaming feature the Grace CPU Superchip, BlueField-3 and NVIDIA A16 GPUs.

The servers are based on four new system designs featuring the Grace CPU or the Grace Hopper Superchip, both of which were announced at previous GTC events. The Grace CPU Superchip features two CPUs connected through Nvidia NVLink C2C interconnect, with up to 144 Arm V9 cores and a 1 TB/s memory subsystem. The Grace Hopper Superchip is a high-powered APU, combining an Nvidia Hopper GPU with a Grace CPU over NVLink C2C in an integrated module designed to address HPC needs.

Discuss on our Facebook page, HERE.

KitGuru Says: What did you all think of Nvidia's Computex keynote this year? 

Become a Patron!