Two AMD EPYC 7742 64-core/128-thread CPUs
At its digital GPU Technology Conference 2020 which kicked off today, Nvidia officially announced the new Amper architecture based GA100 GPU, including the A100 Tensor Core GPU which will be a part of the DGX A100 AI System, and it appears that that same system will be powered by none other than two AMD EPYC CPUs.
GPU compute beast for DGX A100 AI system
The focus of the GTC 2020 was Nvidia's new A100 Tensor Core GPU, which is based on the new Ampere architecture GA100 GPU, and will be a part of Nvidia's new DGX A100 AI system. As a direct successor of the V100 accelerator, the A100 will bring both significant compute performance improvements and new features.
54 billion transistors for 128 SMs or 8,192 CUDA cores and more
Nvidia GTC 2020 has kicked off and, as expected, we got to see the new Amper-architecture with GA100 GPU and A100 Tensor Core GPU in full details.
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.