November 10, 2024. Arm launched its new Cortex-M42 processor, designed to enhance machine learning capabilities in small devices. The Cortex-M42 is important because it allows for on-device AI processing in budget-conscious wearables and industrial control devices, even without access to cloud services. If you're interested in buying a chip that provides effective machine learning intelligence on-device for your battery-powered system, this is something to consider.
New processor enables local AI in wearables and industrial controllers
The Cortex-M42 is a new core based on the Armv8.1-M architecture. It's optimized for running vectorized computations. It incorporates Helium technology to speed up the data processing workloads that have been too complex for microcontrollers of this type. The processor is intended for use in wearables, industrial control, and robotics, where power and size are major considerations.
Specifications
- Architecture: Armv8.1-M
- Vector Extension: Helium
- Vector Compute Performance: Up to 3x compared to Cortex-M33
- Vector Compute Energy Efficiency: Up to 2x compared to Cortex-M33
According to Arm, the Cortex-M42 is capable of providing up to 3x vector compute performance over the Cortex-M33. They're also up to 2x more energy efficient running the vector computations than Cortex-M33. As a result, developers will be able to run more complex DSP algorithms in the smaller, longer-battery-life systems.
Cortex-M42 cores are available through Arm's Flexible Access and Total Access. These licensing arrangements give chipmakers a way to incorporate the core in their own system-on-chip designs. They may or may not require an up-front fee.
Some may have lower royalty rates. The Cortex-M42 joins the Cortex-M4 in the lineup of cores, and it's a specialized processor for AI-capable embedded systems.
Comments (0)