Infrastructure
Enterprises buying GPU cloud capacity for large-scale AI training should expect networking choices like co-packaged optics to shape cluster reliability and efficiency, so they should ask providers about their interconnect design when comparing options.
Cloud provider Lambda is using NVIDIA Photonics co-packaged optics (CPO) switches to connect NVIDIA GB300 NVL72 racks with more than 10,000 GPUs. The network is aimed at making large AI training clusters faster, more efficient and more resilient. The item is an NVIDIA Blog post describing the deployment.
Read the full article at NVIDIA
CoFabrix summarises and comments on this story. The original reporting belongs to NVIDIA.
Find out where your organization stands -- and what to do about it.