IBM has created an upgraded version of its 2021 AI Telum processor to accelerate the development of industrial-scale deep learning models.
Called the Artificial Intelligence Unit (AIU), it is an application-specific integrated circuit that can run any deep learning task, from rapidly processing images to translating spoken languages.
Taking five years to research and develop, the AIU is computational processing hardware that is designed to handle AI calculations requiring matrix and vector multiplication.
Through its plug-in-and-play design, AIU can run and train deep learning models quicker and more efficiently than a CPU.
Speaking on the new chip, IBM said that current CPUs and GPUs can’t keep up with the general-purpose deep learning models that are being developed. Optimised computing has become a necessity for most AI applications that use matrix or vector multiplication operations – a common feature in deep learning development.
Although AI models have become more efficient and accurate in recent years, the hardware they run on still falls short in comparison.
IBM added that while for some AI models edge devices and cloud servers will suffice, there has been a need to invest substantially in new AI hardware platforms for more demanding deep learning models.
In 2019, IBM established its AI Hardware Centre to increase hardware efficiency for AI by 2.5x per year. By 2029, the company aims to train and run AI models one thousand times faster than it could a decade prior.
“Deploying AI to classify cats and dogs in photos is a fun academic exercise. But it won’t solve the pressing problems we face today. For AI to tackle the complexities of the real world — things like predicting the next Hurricane Ian, or whether we’re heading into a recession — we need enterprise-quality, industrial-scale hardware,” IBM said in a statement.
Want to learn more about edge computing from industry leaders? Check out Edge Computing Expo taking place in Amsterdam, California and London.
Explore other upcoming enterprise technology events and webinars powered by TechForge here.