MacMusic  |  PcMusic  |  440 Software  |  440 Forums  |  440TV  |  Zicos
its
Recherche

xAI picked Ethernet over InfiniBand for its H100 Colossus training cluster

mardi 29 octobre 2024, 20:30 , par TheRegister
Work already underway to expand system to 200,000 Nvidia Hopper chips
Unlike most AI training clusters, xAI's Colossus with its 100,000 Nvidia Hopper GPUs doesn't use InfiniBand. Instead, the massive system, which Nvidia bills as the 'world's largest AI supercomputer,' was built using the GPU giant's Spectrum-X Ethernet fabric.…
Lire la suite sur TheRegister
https://go.theregister.com/feed/www.theregister.com/2024/10/29/xai_colossus_networking/

Voir aussi

its Microsoft turning away AI training workloads – inferencing makes better money TheRegister31 Oct
colossus Meta’s Next Llama AI Models Are Training on a GPU Cluster ‘Bigger Than Anything’ Else Wired: Tech.31 Oct
nvidia 6 Prompt Engineering Courses: Free Training for Beginners eWeek28 Oct
infiniband Contractors training Amazon, Meta and Microsoft’s AI systems left without pay after Appen moves to new platform OS News25 Oct
training Intern allegedly messed with ByteDance's LLM training cluster TheRegister22 Oct
ethernet Planned Adobe app to put an end to AI training on images and clips ComputerWorld 9 Oct
over Security awareness training is not enough to stop breaches BetaNews 9 Oct
picked Meta Confirms It Will Use Ray-Ban Smart Glasses Images for AI Training Slashdot 3 Oct
hopper Zyxel launches Wi-Fi 7 gateways with advanced 10G Ethernet capabilities BetaNews 3 Oct
xai Meta Hit With New Author Copyright Lawsuit Over AI Training Slashdot 3 Oct
cluster This Startup Wants YouTube Creators to Get Paid for AI Training Data Wired: Tech.30 Sep
system OpenAI to reveal secret training data in copyright case – for lawyers' eyes only TheRegister26 Sep
News copyright owned by their original publishers | Copyright © 2004 - 2024 Zicos / 440Network
Date Actuelle
jeu. 31 oct. - 11:23 CET