A staff of engineers at AI inference know-how firm BitEnergy AI stories a technique to cut back the vitality wants of AI functions by 95%. The group has printed a paper describing their new method on the arXiv preprint server.
As AI functions have gone mainstream, their use has risen dramatically, resulting in a notable rise in vitality wants and prices. LLMs corresponding to ChatGPT require lots of computing energy, which in flip means lots of electrical energy is required to run them.
As only one instance, ChatGPT now requires roughly 564 MWh every day, or sufficient to energy 18,000 American houses. Because the science continues to advance and such apps grow to be extra widespread, critics have steered that AI functions is likely to be utilizing round 100 TWh yearly in just some years, on par with Bitcoin mining operations.
On this new effort, the staff at BitEnergy AI claims that they’ve discovered a option to dramatically scale back the quantity of computing required to run AI apps that doesn’t lead to diminished efficiency.
The brand new method is fundamental—as a substitute of utilizing advanced floating-point multiplication (FPM), the strategy makes use of integer addition. Apps use FPM to deal with extraordinarily massive or small numbers, permitting functions to hold out calculations utilizing them with excessive precision. It is usually probably the most energy-intensive a part of AI quantity crunching.
The researchers name their new methodology Linear-Complexity Multiplication—it really works by approximating FPMs utilizing integer addition. They declare that testing, up to now, has proven that the brand new strategy reduces electrical energy demand by 95%.
The one downside it has is that it requires completely different {hardware} than that at the moment in use. However the analysis staff additionally notes that the brand new kind of {hardware} has already been designed, constructed and examined.
How such {hardware} could be licensed, nonetheless, continues to be unclear—at the moment, GPU maker Nvidia dominates the AI {hardware} market. How they reply to this new know-how might have a serious affect on the tempo at which it’s adopted—if the corporate’s claims are verified.
Extra data:
Hongyin Luo et al, Addition is All You Want for Vitality-efficient Language Fashions, arXiv (2024). DOI: 10.48550/arxiv.2410.00907
arXiv
© 2024 Science X Community
Quotation:
Integer addition algorithm might scale back vitality wants of AI by 95% (2024, October 12)
retrieved 12 October 2024
from https://techxplore.com/information/2024-10-integer-addition-algorithm-energy-ai.html
This doc is topic to copyright. Aside from any honest dealing for the aim of personal examine or analysis, no
half could also be reproduced with out the written permission. The content material is offered for data functions solely.