NXP Aims for Cutting-Edge AI with New Microcontroller Families


This article is part of our Embedded World 2022 coverage.

NXP hopes to bring artificial intelligence (AI) to low battery embedded and peripheral devices with a new family of 32-bit Arm Cortex-M microcontrollers that contain a custom neural processing unit (NPU) for the first time.

The flagship MCU in the company’s new MCX portfolio uses the NPU to run machine learning workloads 30 times faster than a CPU alone, bringing AI out of the cloud and into consumer and industrial IoT devices.

NXP said the Cortex-M33 microcontrollers span multiple families, each covering a different base in the world of real-time embedded systems, giving engineers more options to select the component that’s right for them. These include the high performance and highly secure MCX “N” family, the low cost MCX “A” range, the MCX W low power wireless connectivity and the ultra low power MCX L series for powered devices. by battery.

The MCX series is supported by NXP’s popular suite of MCUXpresso development tools and software. The company said the unified software suite helps maximize software reuse to speed development.

NXP emphasized that the MCX portfolio does not replace the more than 1,000 SKUs of its LPC and Kinetis 32-bit microcontroller lines, and plans to continue to provide and support them in the future.

“With the MCX portfolio, we select the best LPC and Kinetis peripherals and architectures, bringing them together in next-generation microcontrollers,” said CK Phua, Microcontroller Product Manager at NXP.

Flexibility in a nutshell

At the hardware level, all edge computing devices look alike. But each device is unique in its own way.

Today, engineers must navigate a complex landscape of performance requirements, as well as wired and wireless connectivity options, while balancing system security, energy efficiency, and cost.

“It’s not a unique situation,” Phua said. Ideally, you would want to be able to revise the underlying hardware as market needs change without having to revise the software running on it.

To bring more flexibility to the table, NXP isn’t just adding more SKUs to its MCU portfolio. “We are entering a new era of edge computing, which requires us to fundamentally rethink how best to design a portfolio of microcontrollers that are flexible, scalable, optimized, and can form the basis of energy-efficient industrial and IoT applications today. today and in the decades to come,” said Ron Martino, vice president and general manager of edge processing.

With each 40nm microcontroller in the family sharing the same CPU architecture, based on Arm’s Cortex-M33, NXP tries to make it as easy to scale to different devices as possible and to maximize software reuse.

MCX microcontrollers feature up to 4MB of on-chip flash memory, power-efficient cache and advanced memory management controllers, and up to 1MB of on-chip SRAM with error-correcting code (ECC) to improve the real time. performance.

The MCX line is based on NXP’s security-by-design approach, offering secure boot with an immutable root of trust, hardware-accelerated cryptography and, on select products, its EdgeLock secure subsystem. Flexible interfaces and smart peripherals give developers greater design flexibility.

Dial “N” for NPU

The flagship MCX “N” family features the same Cortex-M3 processor as other MCUs, giving it what NXP has called the best combination of performance and power efficiency for running real-time smart edge devices.

The MCX N has many of the same building blocks that helped put the company’s 32-bit MCUs on the map, which have shipped in billions of units to date. But it also contains the first NPU designed by NXP. The NPU is able to run machine learning on the device itself rather than in the cloud, reducing latency that is intolerable for real-time edge devices.

The MCX N has the same Cortex-M33 processor as the other members of the family, clocked between 150 and 250 MHz. It is complemented by a range of other peripherals, including a digital signal processing (DSP) subsystem.

NXP is betting more on the MCX N family NPU as it believes that more embedded and peripheral devices powered by 32-bit microcontrollers will benefit from machine learning in the coming years.

This is reflected in its decision to keep machine learning engine development in-house instead of buying Ethos NPU plans from Arm. NXP first announced its ambitions around on-device AI in 2018.

Bringing chip development in-house also gives it more control over the future NPU roadmap and the flexibility to adjust it for customers. NXP said the NPU is scalable from the 32 operations per cycle in the NPU at the core of the MCX N family to over 2,000 operations per cycle. A future extension is possible. NXP plans to connect the NPU to other parts of its in-vehicle portfolio, including its i.MX RT and i.MX lines.

“The NPU architecture will be scalable,” Phua said. “We will cover all CPU architectures in NXP, but all with a single architecture for the NPU.”

The NPU inside the MCX N can pump out up to 8 billion operations per second at 250 MHz, a fraction of the performance of AI engines designed by Apple, MediaTek and Qualcomm for mobile phones.

NXP said the MCX portfolio NPU will provide a performance boost to space- and power-constrained IoT devices that run on 32-bit microcontrollers. This opens the door to functions such as classifying objects in video and images and identifying keys in audio. Predictive maintenance in factories is also possible, NXP said.

Machine learning will be supported by NXP’s eIQ ML software development toolkit. Developers can rely on the easy-to-use tools to train models and deploy them to the NPU or CPU, according to the company.

From point “A” to point “N”

On the other side of the portfolio is the MCX A family for embedded devices where cost and time to market are the priority. Designed to be easy to use, the MCX A series runs the same Cortex-M33 as the rest of the family, but with more limited features to keep costs down, including clock speeds from 48 to 96 MHz. It comes with a single-pin power supply, built-in timers, and reduced pin count for a “limited cost” devices.

The MCX A fits into the same enclosure as the MCX N, allowing customers who require more performance, additional memory, or enhanced security measures to use the MCX N as a replacement.

The MCX N also stands out by implementing a secure and immutable subsystem called EdgeLock, NXP says. The so-called secure enclave supports device “root of trust” to perform secure boot, advanced key management, device attestation and trust provisioning to thwart all kinds of attacks on devices IoT.

As noted, the MCX wallet will be supported by the MCUXpresso tool and software suite, the same used by its Kinetis and other 32-bit microcontrollers. This allows developers to reuse large chunks of software from device to device, instead of starting from scratch when designing a new product or upgrading the underlying hardware.

The breadth of the MCX portfolio allows engineers to choose the microcontrollers that best meet their design needs. NXP said maximizing software reuse frees them to invest in differentiating aspects of their application.

“L” for low power, “W” for wireless

The MCX L, based on a Cortex-M33 MCU clocked at 50 to 100 MHz with an optional 50% boost mode, is intended to minimize power consumption in IoT edge devices as much as possible.

The MCX L operates on sub-threshold voltages and uses technologies such as dynamic voltage adjustment and body biasing to reduce dynamic (active) and standby (leakage) power consumption.

NXP said the MCX L series will help extend battery life in industrial and consumer IoT systems and other edge devices where battery power is a precious commodity.

The MCX “W”, with clock speeds of 32-150 MHz, promises to add wireless connectivity to IoT devices with support for Bluetooth LE 5.2 along with a high degree of on-chip integration that reduces BOM costs.

The first MCUs in the MCX portfolio will begin sampling by the end of this year, with volume production expected in 2023.

The company declined to discuss pricing.

Check out more articles/videos in our Embedded World 2022 issue.


Comments are closed.