The Future of Edge Al Multicore Technology Drives AI at the Edge

The 32-bit architecture changed the trajectory of the embedded market.

- By Jim Beneke

Arm's Cortex-M3 was the first to target the microcontroller product line, quickly followed by the -MO and -M4. That happened around 20 years ago and now almost all microcontroller (MCU) vendors are Arm licensees, even those with their own in-house 32-bit cores.

The increased performance this architecture enabled created more customer demand, and this virtuous cycle helped open the market to processors such as the x86 architecture. We're now seeing a subtle but unmistakable shift toward the RISC-V open-source instruction set architecture. Even without Arm’s influence, there’s little doubt that the industry would now favor 32-bit architectures over 8- and 16-bit (although both are still hugely popular).

Meeting the need for complexity in embedded products has always been cited as the reason behind this evolution. More powerful 32-bit architectures support more capable software, including real-time operating systems and, of course, Linux.

The latest development that cannot go unnoticed is the introduction of machine learning (ML) and artificial intelligence (AI) in smaller devices. Edge AI will set the direction of embedded processors for the next decade. While incumbents are evolving to capture this opportunity, there are also many startups hoping to become the new disruptor.

THE ADVANTAGES OF MULTICORE SOLUTIONS

Process technology has developed in parallel to the adoption of 32-bit buses. We have seen continued advancements in both, culminating in MCUs that measure millimeters on each side capable of controlling disproportionately larger systems running highly complex software.