STMicroelectronics launches new edge AI processor series
STMicroelectronics has launched a new generation of 64-bit microprocessors for edge computing. These microprocessors provide improved performance, enhanced security and specialized features that aid real-time processing of data-intensive machine learning algorithms used in industrial applications.
The first processor released is the STM32MP25 microprocessor, which has a specialized neural processing unit to aid in edge AI applications. This unit provides up to 1.35 TOPS of computing power. It is suitable for machine vision and predictive maintenance applications in industrial manufacturing plants.
“The new STM32MP2 Series devices further our investment in application processors, combining 64-bit cores with edge AI acceleration, advanced multimedia features, graphics processing, and digital connectivity,” says Ricardo De Sa Earp, the executive vice president and general manager of General-Purpose Microcontrollers Sub-Group at STMicroelectronics.
The STM32MP25 contains both the Arm Cortex-A35 processor core and the Arm Cortex-M33 embedded core, enabling real-time edge data processing with low latency. Further, the microprocessor features integrated support for 32-bit DDR4 and LPDDR4 memory, allowing for adaptable memory configurations tailored to an application’s specific requirements.
To ensure compatibility with industrial environments, the microprocessor module must adhere to specific standards for critical communication. The microprocessor is equipped with a two-port Gigabit Ethernet TSN switch and PCIe and CAN-FD peripherals. These connectivity options are crucial to the industrial application ecosystem.
STMicroelectronics states that the STM32MP25 microprocessor can capture video data from a 5-megapixel sensor at 30 frames per second. The company says this feature showcases the MPU’s capability to analyze the acquired video using the edge computing accelerator. The module has a 3D graphics processing unit that can handle graphics-intensive tasks like rendering complex visual content while supporting a maximum resolution of 1080 pixels.
According to the company, the microprocessor is also secure with SESIP level 3 certification. It includes Arm’s TrustZone architecture and a resource isolation framework (RIF) to enhance security. The RIF provides isolation of system resources to prevent unauthorized access between different modules. This feature protects critical resources and guards against security breaches.
STMicroelectronics recently made headlines for launching software toolchains and libraries that allow for the programming of intelligent sensor processing units. These units are used for edge computing applications in healthcare, manufacturing and robotics.
Agile Analog unveils chip tech for RISC-V in battery-powered IoT systems
Article Topics
AI/ML | ARM | computer vision | edge AI | IIoT | Industry 4.0 | STMicroelectronics
Comments