Embedded systems and their applications are rapidly diversifying, as are the processors that serve them. As a result, embedded processors are becoming increasingly complex, providing engineers with more and more choices that often overlap in functionality. Although we always welcome more choices, exploring various possibilities may take a lot of time. In order to succeed in dynamic environments, developers need a method to quickly evaluate multiple chip choices from microcontroller units (MCUs) to microprocessor units (MPUs), while simplifying the prototype development process.
One way to assist designers is to adopt a modular approach to hardware processing. By combining a simplified development board with a rich library of extension modules and supporting software, designers can match them as needed.
This article reviews the changes in design requirements for embedded systems and their implications for processor selection. Then explain how NXP's platform helps designers explore multiple processor categories, ranging from low-power MCUs to highly integrated Linux grade MPUs and application processors.
The boundaries of embedded design are becoming increasingly blurred
Until recently, most embedded applications fit into well-defined categories. Simple input/output and control logic belong to the category of 8-bit MCUs; 32-bit MCUs are responsible for handling complex real-time tasks. Application software that requires a complete operating system (OS) or graphical user interface (GUI) belongs entirely to the field of MPU.
Nowadays, these boundaries have become blurred. Due to the addition of complex connectivity features in previous standalone applications, many traditional 8-bit applications have been pushed into the 32-bit domain. The complex software stack is rapidly increasing in real-time applications, integrating the requirements of MCU and MPU applications. At the same time, artificial intelligence (AI) and machine learning (ML) are being integrated into an increasingly wide range of applications.
The distinction between processor categories has also become blurred. High performance MCUs now have graphics accelerators and AI/ML capabilities, which were once patented for high-end MPUs. MPU has added real-time functionality that was previously exclusive to MCUs. In addition, the demand for high-end graphics, AI, and other complex functions has also driven the launch of application processors, whose architecture draws on the design of mobile phones.
All of this is happening in the context of accelerating innovation. From the beginning of the design cycle to the product launch, market dynamics will significantly change project requirements. For example, designs based on headless MCUs may unexpectedly be equipped with touch screens, requiring an upgrade to MPU. On the contrary, the product marketing team may decide at the last minute that high-end products need to be equipped with entry-level versions, thus rushing to find lower cost processors.
These trends and changes have spurred the need for a processor evaluation ecosystem so that designers can easily explore various options. Traditional evaluation boards are unable to meet this demand. They are often designed to showcase all the important features of a processor family, often using complex designs optimized for narrow application ranges. Therefore, the effort invested in one evaluation board is rarely translated into another evaluation board.