MCUs and MPUs, which does your project really need?

The basics of microcontroller units (MCUs) and microprocessor units (MPUs) haven’t changed too much. But application requirements have. What was once a simple automotive, industrial, or medical control system may now include edge AI processing, increasingly demanding security requirements, or complex 3D animated human machine interfaces (HMIs). That can change the game when choosing between MCUs and MPUs.

Today, MCUs are more widely used than MPUs, but that’s also beginning to change. MCUs are optimized for low power, IoT, and sensor applications. MPUs are capable of handling high-definition (HD) HMIs, machine learning, machine vision, and other computationally demanding applications (Figure 1).

Figure 1. These traditional differences between MCUs and MPUs are blurring. (Image: ST)

For the lowest power consumption, MCUs are still preferred, and for high performance, MPUs are dominant. But that’s changing. In response to diversifying application needs, MCUs, MPUs, and system architectures are evolving.

MCUs are getting more powerful and capable. The clock speeds of 32-bit and 64-bit MCUs have increased, also increasing power consumption. In addition, some MCUs have multi-core architectures and can run real-time operating systems (RTOS) to manage complex tasks.

In addition to high computational power and their ability to interface with external components, MPUs are becoming more integrated and power-efficient, challenging MCUs in some embedded applications like edge sensor nodes that can benefit from local data processing to support increased system performance. There are MPUs that have been optimized for use in battery-powered applications, once the sole preserve of MCUs.

Smart home applications provide some good examples. Smart home hubs that need significant data processing, multitasking, and integration with various protocols often use MPUs. Advanced consumer appliances with graphical displays and connectivity also leverage MPU capabilities.

On the other hand, in basic smart refrigerators, washing machines, thermostats, and ovens, MCUs can provide the needed control functions and some advanced features, providing a cost-effective solution with sufficient processing power for dedicated tasks.

Figure 2. These traditional structural differences are being blurred by specialized MCUs with integrated PCIe interfaces and other advanced features. (Image: Northwest Engineering Solutions)

Not an either-or choice

Depending on the system requirements, a hybrid solution that includes both an MCU and an MPU can produce maximum benefits. A low-power MCU can be used for handling real-time control, with the MPU taking on the tasks that require higher performance. For example, applications with diverse requirements, including the need to balance high-level features, like connectivity, complex algorithms, and GUIs, with the strict timing requirements of embedded control, may benefit from a hybrid approach.

Devices like smart home hubs or advanced security cameras that require both the processing power for features like AI or video analysis and the ability to manage sensors and actuators in real-time are examples where a hybrid approach may be beneficial.

So are many industrial or vehicle motion control applications where MPU can run the operator’s display and remote diagnostics, while an MCU handles motor speed, sensor readings, and safety interlocks.

SoCs can change the equation

SoCs can provide designers with another way forward for advanced systems. That can include adding functions like wireless connectivity, graphic processing units (GPUs), and even digital signal processors (DSPs) into a single package.

The synergy of using different cores and functions enhances performance while constraining solution sizes. For example, most smartphones have a SoC that uses an MPU for running the OS and an MCU for power management, display control, and other functions.

For IoT devices, the combination of an MCU and a wireless module like Bluetooth or Wi-Fi, plus the needed antenna and other glue components, can support cost-effective, compact, and easy-to-implement solutions. Many of these wireless MCUs are pre-certified for operation in multiple countries, further speeding time to market.

Summary

Designers can’t rely on yesterday’s distinctions between MCUs and MPUs when making decisions regarding modern applications. The growing variety of applications that require both control functions and data processing is expanding. That has blurred the line between MCUs optimized for control and MPUs optimized for data processing. MCUs are available with capabilities formally relegated to MPUs and some MPUs can handle functions once the domain of MCUs. If that’s not complex enough, designers also have to consider the emergence of SoCs when identifying the optimal platform for a given application.