Training Large LLM Models With Billions To Trillion Parameters On ORNL’s Frontier Supercomputer


A technical paper titled “Optimizing Distributed Training on Frontier for Large Language Models” was published by researchers at Oak Ridge National Laboratory (ORNL) and Universite Paris-Saclay. Abstract: "Large language models (LLMs) have demonstrated remarkable success as foundational models, benefiting various downstream applications through fine-tuning. Recent studies on loss scaling ... » read more

Chip Industry Week In Review


By Jesse Allen, Karen Heyman, and Liz Allan Japan's Rapidus and the University of Tokyo are teaming up with France's Leti to meet its previously announced mass production goal of 2nm chips by 2027, and chips in the 1nm range in the 2030s. Rapidus was formed in 2022 with the support of eight Japanese companies — Sony, Kioxia, Denso, NEC, NTT, SoftBank, Toyota, and Mitsubishi's banking arm, ... » read more