Page 3 - Can AI take education to a new level

ANALYSIS

Home

Chip sector gears up for AI revolution
News

How the chip sector
is gearing up for While Nvidia establishes its AI dominance, Intel works on AI chips running on supercomputers. Cliff Saran reports
the AI revolution

How ISC2 aims
to overcome he most recent quarterly earnings from Meta and Google mainly due to the development of large language models (LLMs)
cyber barriers demonstrate the trend among hyperscalers to extend the and generative AI. This has led to growth of 195% in its datacen-
Tlife of datacentre servers. tre compute business.
Editor’s comment Google announced that it saved almost $3bn in nine months The chip company is working with HPE’s Cray supercomputer
by extending server life to six years, while Meta’s chief financial division on a family of artificial intelligence-optimised supercom-

Buyer’s guide officer, Susan Li, revealed the company could see little perfor- puters based on its Grace Hopper GH200 Superchips. HPE said
to the future of mance gain from new server chips. Instead, she said: “We’re the combination of AI software and hardware offers organisa-
business software
still seeing strong performance gains for a new GPU generation, tions the scale and performance required for big AI workloads,
so those will depreciate on a relatively shorter timeframe and such as training large language models and deep learning
Harnessing large
language models we’ll just continue to evaluate how to most efficiently use the recommendation models.
for education CPUs and GPUs across our fleet.” According to HPE, using its Machine Learning Development
Environment on this system, the open source 70 billion parameter
AdvAncing Ai options Llama 2 model was fine-tuned in less than three minutes. Overall
Recognising the change in buying strategy among the hyperscal- system performance improved by 200% to 300%.
ers and in anticipation of a shift to artificial intelligence (AI) work- Rival chipmaker Intel reported a 10% decline in its datacentre
load optimisation among enterprise customers, semiconductor and AI business in the third quarter of 2023. Keen to plug this
manufacturers are beginning to ramp up their AI portfolios. gap, Intel has begun ramping up its AI efforts.
Nvidia’s share price and strong financial results show how the At SC23, the international conference for high-performance
GPU company has pivoted to becoming the leading supplier of computing (HPC), networking, storage and analysis, it unveiled
AI-accelerated hardware. work on a new supercomputer, Aurora, with Argonne National
Its second quarter earnings saw strong demand for the Nvidia Laboratory and industry partners to create what it describes as
HGX platform based on its Hopper and Ampere GPU architectures, “state-of-the-art foundational AI models for science”.

computerweekly.com 21-27 November 2023 3