Remove AI Networking Remove Ethernet Remove Protocol
article thumbnail

HN728: How Drivenets Leverages Ethernet Fabrics For AI Networking (Sponsored)

Packet Pushers

To run AI workloads, a network needs thousands of GPUs and those GPUs must operate in sync. While there are advantages to using Ethernet for AI networking (including engineers well-trained in the protocol and a robust ecosystem), it wasnt designed. Read more »

article thumbnail

A RoCE network for distributed AI training at scale

Engineering at Meta

AI networks play an important role in interconnecting tens of thousands of GPUs together, forming the foundational infrastructure for training, enabling large models with hundreds of billions of parameters such as LLAMA 3.1 The growing prevalence of AI has introduced a new era of communication demands.

Network 132