With a Focus on Industry 4.0, ST Unveils Second Series of STM32 MPUs
The new family of MPUs offers some unique architectural features for industrial applications.
The industrial space is a broad and wide-reaching field, encompassing office buildings, factories, and everything in between. Despite this diversity, all smart industrial applications require high-performance computing and processing. This demand for industrial computing is only expected to increase as Industry 4.0 takes hold.
ST's new microprocessors are built to facilitate smart industrial designs. Image courtesy of STMicroelectronics
Recently, STMicroelectronics announced a new Industry 4.0-specific processor family, called the STM32MP2, as the second generation of the STM32 family. This new family includes a new architecture said to improve performance and efficiency. In this piece, we’ll take a deeper look at the STMP32MP25, the first offering in the new STM32MP2 family, to understand the new architecture and its value for industrial computing applications.
STM32MP2: Basic Compute
Breaking down the STM32MP25, the first thing we can look at is the system’s computing capabilities.
Unlike its STM32 predecessors, which are 32-bit solutions, the STM32MP25 is a 64-bit industrial solution. With this increase in bit width, the STM32MP25 can process larger numbers, offer more memory addresses, and include larger registers than a 32-bit solution.
Simplified block diagram of the new STM32MP2 family. Image courtesy of STMicroelectronics
The new processor is built around either single or dual Arm Cortex-A35 cores, which can run clock speeds up to 1.5 GHz, as well as a single Arm Cortex-M33 core clocked up to 400 MHz. With Arm Cortex-A35 cores designed for greater IoT efficiency and Cortex-M33 designed for real-time digital signal processing, it’s clear that ST's focus here is a balanced solution that blends both power efficiency with fast processing. With this central computing system, the new architecture is said to achieve up to 6,000 DMIPS of compute.
Acceleration for High AI Performance
To complement the STM32P25’s main computing system, the new architecture also incorporates some hardware acceleration blocks.
Specifically, the system features a neural processing unit (NPU), of which not much is currently known, but is said to achieve AI performance of up to 1.35 TOPS. ST claims its NPU is optimized for applications such as machine vision and predictive maintenance. To further support machine vision, the system features a MIPI CSI-2 camera interface with Lite-ISP—a feature that may simplify the system design of high-performance cameras.
For multimedia processing needs, the system features a 3D graphics processing unit (GPU). The GPU can support 1080p resolution for graphics and video and is also designed for driving user interfaces with real-time responsiveness.
Broad Connectivity Support
Finally, a good Industry 4.0 processor needs to communicate effectively with the rest of the system. To support this, ST incorporated a number of communication interfaces into its STM32MP2 family.
Key features of the STM32P2 family. Image courtesy of STMicroelectronics
The STM32P25 supports gigabit time-sensitive networking (TSN) (including a two-port gigabit Ethernet TSN switch), as well as PCIe, USB 3.0, and CAN-FD. This wide-ranging support for a plethora of high-speed interfaces, as well as TSN, is crucial for Industry 4.0 solutions where end nodes may require different communication protocols.
To further increase compute capabilities, the STM32P25 offers interfaces for 32-bit DDR4 and LPDDR4. This helps the system remain scalable should the computing and memory needs of the system increase over time.
A Final Look at the STM32P2
When we dissect the new STM32P2 architecture, it’s clear that these devices were indeed designed with smart industrial applications in mind. Between efficient and powerful computing cores, machine learning and graphics acceleration capabilities, and several communication interfaces, the STM32P25 may be a useful choice for industrial designs that rely heavily on communicating within a larger system.