ARM's new chip architecture promises a big boost for Artificial Intelligence

From Siri to Netflix and Alexa to Spotify, we all use various forms of Artificial Intelligence on a daily basis. However, the current generation of chips and processors used to power AI are operating at near capacity. That’s why the big chip designers and manufacturers are releasing new architecture to accommodate tomorrow’s technology. At the moment, a large proportion of AI processing takes place in the cloud, partly because the chips in mainstream production today are unable to handle the sheer amount of data and power required to process the algorithms and data streams. This information that gets handled remotely is then transferred via the internet. In a perfect world, this isn’t a problem. However, in reality, we have to contend with dodgy WiFi networks and cellular connections that drop off a cliff whenever you pass through a bridge. This poses a problem for the technologies of the future as a strong, secure and continuous connection to the cloud cannot always be guaranteed. Self-driving cars, for example, will need processing to be conducted at a local level as any interruption could bring about catastrophic consequences. To meet the needs of these demands, chip designers have gone back to the drawing board. ARM, the company that designs the chips used in most major smartphones, is looking to address this by launching a major update. Entitled DynamIQ, this upgrade will boost performance in the field of AI by 50 times when compared to its current chips. Previously, ARM has opted to focus its attention on creating chips that are extremely power efficient. This approach has helped the company succeed in locking down the smartphone and mobile devices market, an area where its competitors failed to produce as effective components. But given the adoption of AI-features by smartphone manufacturers, and the way that technology is heading, ARM has opted to evolve. Believed to be released next year, ARM’s new central processing units (CPU) will feature multiple processing cores clustered together, up to a maximum of eight. This is an increase from the maximum of four that can be used at the moment and, in some instances, there will also be the inclusion – in some instances – of a processor for handling AI-related algorithms. Every new ARM chip design - dubbed the Cortex-A series - will be based on this new architecture. ARM is not the only chipmaker to release a major update to its existing architecture. Intel has committed to delivering gains of 15% or more with every new generation, and AMD has announced its Ryzen series of chips delivers a 40% performance gain in certain metrics. Along with AI, the Internet of Things and self-driving vehicles, other areas such as virtual reality and super-charged gaming devices are also forcing firms to research and release high-performance chips.

For more information on ARM's new chip architecture promises a big boost for Artificial Intelligence talk to Cyclops Electronics Limited