Handbags at dawn

AMD has come out with a benchmark of its own between its Instinct MI300X and Nvidia’s H100 (Hopper) GPUs.

The DGX GH200 AI Supercomputer makes an appearance

Nvidia kicked off its Computex 2023 presence with a keynote led by its founder and CEO, Jensen Huang, and a bunch of big announcements, starting with the full production of its Nvidia GH200 Grace Hopper Superchip, which will power systems from its various partners as it becomes available later this year. Nvidia also unveiled the massive 1-Exaflop DGX GH200 AI Supercomputer.

Brings 16,896 CUDA cores and up to 80GB of HBM3 memory

During the GTC 2022 keynote, Nvidia has unveiled its latest H100 accelerator, a beast that uses 4nm GH100 GPU and packs 80 billion transistors, 4th gen Tensor cores, and promises some impressive performance gains thanks to the new Hopper architecture.

Rabbiting on

The GPU maker named after a Roman vengeance daemon, Nvidia has announced the release of its Hopper GPU architecture.

If only it can make this pesky kangaroo go away

Nvidia wants to use the name Hopper for its coming architecture range, the only thing standing in its way is a kangaroo.

TSMC 7nm, next-gen Hopper 5nm

There have been rumors for quite some time about the Nvidia Ampere and the manufacturing nod. Nvidia’s first AI-based solution comes from TSMC, while some GPUs will be manufactured by Samsung too.