NXP’s new chips for edge computing target home, Smart City applications

NXP’s four new series of devices are divided into high-performance MCX N series, cost-optimized and analog-focussed MCX A series, low-power wireless connectivity MCX W series, and low-power MCX L series. The devices are the “first instantiation” of NXP’s specialized neural processing unit for accelerating edge inference, capable of delivering up to 30x better machine learning compared to a CPU performance.

The MCX portfolio is designed to solve the challenge of integrating high computational power, improved performance, security features, connectivity options, and balancing the total system cost and energy requirements. Since the devices are based on the existing Arm Cortex-M cores, the new set of four microcontrollers features up to 4MB of on-chip flash memory, low-power cache, and advanced memory management controllers. For the processing memory, the hardware will support up to 1MB of on-chip SRAM to serve the advanced real-time edge applications.

On the software side of the MCX microcontrollers, NXP’s elQ ML software development environment will support machine learning and run-time inference. The software tool provides simplified methods to train machine learning models that target NPU or the CPU core and then deploy them on the desired microcontroller. Integrating security features are most important design parameter for embedded devices that are meant to be deployed at the edge. The four new edge NXP MCX microcontrollers are built on the NXP’s security-by-design approach featuring root-of-trust, hardware accelerated cryptography and built-in EdgeLock secure system.

The new NXP’s MCX microcontroller lineup will be demonstrated at the Embedded World 2022.

Article Topics

ARM  |  chip  |  consumer devices  |  encryption  |  NXP Semiconductor  |  smart city