Though 2014 was relatively a brief time ago, the advancement of technology over the last few years has been astounding. Consider the acceleration of Arm based shipments, which over the last 4 years has doubled. That is an increase of 50 billion devices which integrate an Arm CPU from 2013 to 2017! Back in 2014, the super-scalar Arm® Cortex®-M7 processor was unleashed into a market poised for the age where smart and connected things would proliferate into the billions. Today, with 8.4 billion connected things having shipped in 2017, the capabilities of the top tier Arm Cortex-M processor has already enabled many classes of embedded devices, from wearables to industrial controllers. Now with broad industry focus on the application of artificial intelligence, machine learning, vision and voice processing that require more capable edge computation, the Arm Cortex-M7 and the NXP SoC platforms which integrate it, are crossover points for the MCU developer to help achieve the capabilities needed to address emerging requirements.
Central to the compute capability of the SoC is the CPU. Close to the release of the Arm Cortex-M7, I had the pleasure of collaborating with Arm on a whitepaper that explores the processor capabilities in detail. The information there is arguably more relevant today, so please have a read. At the time we wrote the whitepaper, little did we know the extent to which we could achieve with the processing speeds and peripherals that we’ve done with the new i.MX RT. There is an excellent whitepaper that outlines how its attributes make it unclassifiable in the traditional embedded space. With the i.MX RT family, we have created the new ‘crossover processor’ class. One of the benefits of this class of devices is shown in the below graph which relates performance to the price of different options in the market.
So, I would like to add the following addendum to the whitepaper to explore the Cortex-M7 integration for our i.MX RT crossover processor.
The below diagram captures the architectural details. With regards to cache, the i.MX RT integrates 32KB for the instruction and 32KB for the data caches. This is the largest size in the market and ensures that the processor is not limited by any delays imposed by slower memories. For the Tightly Coupled Memory (TCM), the i.MX RT has a FlexRAM block of memory. This allows customization of the TCM up to the largest size available (256KB I-TCM, 256KB D-TCM). The user can select this maximum size, or repurpose the FlexRAM to work as On-Chip SRAM.
With regards to the use of AXI on the i.MX RT, there are a broad range of AXI masters which are integrated. Some specific peripherals to highlight which are relevant to emerging trends are the LCD Controller, 2D graphics acceleration engine, Parallel camera sensor interface, and cryptographic accelerator (Data Co-Processor-DCP). These components differentiate the i.MX RT in the market and align to the need for the highest compute performance together with reliable security and assured privacy at the lowest cost possible. Finally, most relevant to the computational capabilities of the i.MX RT is the processor speed. Reaching 600MHz allows the RT to be throttled up for the most intensive calculations.
NXP IoT and Security Solutions team, which applies NXP technology to address market challenges, is focusing on enhancing the i.MX RT with voice and vision components. For the next 100 billion Arm-enabled chips coming in the next 4 years, I expect many to take advantage of the capabilities provided by the CPU and the SoCs.