5 reasons why MIPS M-class CPUs are ideal for IoT

In a recent editorial for the 2016 CES issue of i3 magazine, I wrote about the challenges addressing metropolitan and large urban areas, and how technology can provide some much-needed solutions to the biggest problems facing humankind today.

I also talked about the role of sensors in smart cities and provided a high level overview of the new class of processors needed to power these new devices. In the article below I will try to get more specific and explain why I believe our MIPS M-class CPUs provide a superior CPU solution for the requirements of the IoT market.

There are many factors influencing the choice of a particular processor for a semiconductor device. In the case of the IoT segment, those factors include the hardware characteristics (performance, power, and area), the software support, and the feature set – particularly around security.

If you’d like to find out more about the hardware specifications, I would definitely recommend visiting our MIPS Warrior CPUs webpage. It has all the information you need, including a very detailed analysis of performance, power and area at specific process nodes. Needless to say, our CPUs are highly competitive, delivering faster performance and better power efficiency.

However, picking a CPU for IoT is not all about choosing the best hardware design. Here are five unique ways in which M-class processors stand out from the crowd:

Hardware virtualization support

The M51xx family (the M5150 and M5100 CPUs) implements Release 5 of the MIPS architecture which introduces full hardware virtualization support, among several other new features. M51xx M-class CPUs are the only microcontroller-class processors to deliver full hardware virtualization, supporting up to seven guest VMs (virtual machines) per CPU.

Hardware virtualization support opens up a number of unique use cases for MCUs and enables new and exciting functionality for IoT devices.

OmniShield - security for all devices__f

For example, a MIPS-based chip can implement multiple isolated containers, each running its own operating system or application securely and reliably. In the case of connected home applications, this means separating firmware updates from the code that runs your smart lighting or heating systems. Since the two are kept separate, if the firmware update becomes corrupted at any stage, it won’t break the functionality of your smart device.

A cleaner, more elegant architecture

The MIPS architecture features several distinct strengths that play into IoT use cases.

For example, MIPS CPUs expose 32 general-purpose registers to programmers (double the number available in competing CPUs). In addition, the MIPS architecture specifies additional shadow registers that enable faster context switching when servicing interrupts, leading to faster overall performance in real-time scenarios.

For IoT systems where reducing costs is a top priority, MIPS M-class CPUs also support microMIPS – a set of instructions designed to reduce memory requirements in semiconductor devices. microMIPS includes more efficient memory addressing and achieves more compact code size, helping system designers to cut back on memory specifications and therefore reduce manufacturing costs.

These features also prove really useful for implementing the cryptographic processing in home automation systems. One example is OberonHAP – a fast, tiny and trustworthy implementation of Apple’s HomeKit Accessory Protocol (HAP). OberonHAP has been developed, and is continuously being further improved, by a seasoned Swiss engineering team. In a benchmark available from the company’s website, the MIPS M-class CPU is 20-25% faster per MHz than its closest competitor for a typical HomeKit session.

Cortex-M vs microAptiv for HomeKit IoT

A fully-featured MMU and cache memories

Not only is the MIPS M-class family blazing-fast, but several CPU configurations come equipped with a full MMU (Memory Management Unit) as well as instruction and data cache controllers. M-class processors have the feature set and performance needed to run any RTOS (real-time operating system) as well as a full Linux or Unix-like distribution.


For example, the latest PIC32MZ processor from Microchip Technology has been shown to run 4.4BSD, including a fully-featured networking TCP/IP stack and the Ethernet driver. This capability means that an M-class CPU offers more choice for system architects looking to support rich operating systems and low power connectivity stacks for IoT. In the case of wearables, running Linux on M-class CPUs helps developers create far more sophisticated UIs to ones we have today on microcontroller-based watches or fitness bands.

Better DSP performance for audio processing

Many in the industry see voice (pardon the pun) as one of the best ways of interacting with IoT devices; several connected home products today already implement the voice control paradigm extensively.

This is why we’ve equipped our M-class processors with the absolute best DSP instruction set available in the industry, delivering 2x the performance of competing CPUs. The chart below shows how our MIPS M-class CPUs fare against another licensable microcontroller-class processor when running typical DSP workloads:

MIPS microAptiv v the competition

In addition, we’ve built an entire IoT IP subsystem optimized around voice processing that system architects can quickly integrate in their SoC design. This subsystem is only one of the many scalable platforms we have available for different IoT devices (more details below).

Easier system integration and cloud connectivity

Silicon vendors can quickly combine our MIPS, PowerVR and Ensigma IP to design next-generation SoCs for connected devices. These IP platforms are designed to accelerate time to market by delivering a truly secure and connected processor solution for consumer and industrial IoT.

Our Ensigma Whisper low-power architecture is extremely configurable and supports a multitude of connectivity standards, including Wi-Fi, Bluetooth Classic, Bluetooth Smart, NFC, and GNSS.

Consumer IoT - MIPS CPU Ensigma RPU

In addition, we also provide a cloud platform called FlowCloud that enables companies to build and manage all aspects of connected services between devices and humans. FlowCloud is designed to address the needs of the consumer and enterprise IoT market, enabling easy product registration and updates as well as access to comprehensive and application-independent services.

If you want to experience a real-world example of how our IoT IP platforms and FlowCloud solution work together, look no further than our Creator Ci40 dev kit.

06 - Creator Ci40 IoT kit - FlowCloud detailed

Creator Ci40 includes 6LoWPAN low-power sensors powered by a MIPS MCU and a high-performance hub integrating a MIPS CPU and Ensigma RPU. Instead of locking developers into a proprietary and closed operating system, the kit is entirely based on open source software technologies (Contiki and Linux, respectively) and stays connected to the cloud via FlowCloud.

Get in touch if you want to learn more

If you want to learn more about how we can help you build a superior IoT product or if you are building an embedded system and simply want to learn more about the MIPS architecture, head over to our contact page and get in touch with our CPU experts.

For more news and updates on MIPS, follow us on Twitter (@ImaginationPR, @MIPSGuru, @MIPSdev, @prpl_foundation), LinkedIn, Facebook and Google+.

Leave a Comment