AIoT: Solutions from Several Semiconductor Manufacturers

This article shares some chip solutions from semiconductor manufacturers in the field of AIoT.

Arm® Cortex®-M52

Arm launched the Arm® Cortex®-M52 processor designed specifically for Artificial Intelligence Internet of Things (AIoT) applications in November 2023. The new Cortex CPU core adds DSP support and ML acceleration (Helium) options.

Helium technology is a vector processing technology configured in the Cortex-M processor series, primarily used for AI computation, and is an extension of the Armv8.1-M architecture. Currently, Cortex-M55 and Cortex-M85 support Helium technology. Its advantages include: replacing traditional DSPs, reducing system costs and software development processes, and achieving machine learning capabilities without a dedicated NPU.

Cortex-M52 provides a simplified migration path from Cortex-M33 and Cortex-M4, meeting a wide range of AIoT application needs.

Compared to other Cortex-M processors supporting Helium technology, the Cortex-M52 mainly makes trade-offs in power consumption and performance cost.

Arm Cortex-M52 Specifications

NXP: Integrating NPU into General-Purpose MCU

NXP has launched a portfolio of MCU products that integrate AI capabilities – MCX.

Among them, the MCX N series is a high-performance, low-power microcontroller with intelligent peripherals and accelerators, including NXP’s elQ Neutron neural processing unit (NPU), suitable for machine learning applications.

The latest products in the MCX N series are the MCX N94x and MCX N54x, which adopt an efficient multi-core architecture and integrate dedicated on-chip neural processing units (NPU).

​

ST STM32MP2: 64-bit Microprocessor with Neural Processing Unit

In May 2023, ST launched its latest 64-bit microcontroller STM32MP2, aimed at Industry 4.0 and advanced edge computing applications, supporting smart interconnected applications.

Main features include:

One Arm Cortex-A53 CPU

One Arm Cortex-M33 CPU

GPU – An unnamed 3D GPU for “advanced HMI”, supporting Vulkan API

AI accelerator – 1.35 TOPS NPU (Neural Processing Unit)

Main application areas:

Industry 4.0, machine learning, smart home, smart metering, payment terminals, Internet of Things (IoT), and edge computing, etc.

Reference:

https://www.nxp.com/company/blog/push-the-edge-of-whats-possible-with-nxps-mcx-n-advanced-series-mcus:BL-NXPS-MCX-N-ADVANCED-SERIES-MCUS

Thank you for reading, and please like and support!