Entering the new era of artificial intelligence (AI), the process of intelligent terminal equipment has accelerated significantly, and a large number of Microcontrollers (MCUs) integrated with neural processing units (NPU) have emerged. As a leader in the MCU industry, STMicroelectronics recently launched the STM32N6 series, filling the gap in its AI product line and integrating into the competitive landscape of this new era.
Source from STMicroelectronics
News of the STM32N6 series MCU came out as early as 2022. Although it was still in the research and development stage at that time, its unique features have attracted much attention in the industry. This much-anticipated product is finally here, with comprehensive specifications that achieve the performance promised in the early stages of development.
STM32N6 Series MCU With AI Performance
Regarding AI performance, the STM32N6 series is equipped with ST's independently developed Neural-ART NPU, which can provide a strong AI computing power of 0.6 TOPS at a frequency of up to 1GHz. This level of computing power improves its machine-learning capabilities by a full 600 times compared to previous top-level STM32 devices.
With its excellent neural network reasoning capabilities, the STM32N6 series easily meets the performance requirements of various computer vision and audio applications, whether in the smart industry, smart home, smart city, automobile manufacturing, consumer electronics, or other intelligent application scenarios.
The advantages of the STM32N6 series microcontroller are not just AI computing power. As an ST new product line, it performs well in conventional performance. The built-in Arm Cortex-M55 core has a main frequency of up to 0.8GHz. It has an unprecedented 4.2MB RAM and integrates an ISP (image signal processing) module that can directly process signals.
It is reported that the STM32N6 series is built based on the advanced 16nm process. Thanks to this advanced process, such a rich configuration can be integrated. With the continuous development of advanced methods, it will become a general trend for MCUs to integrate AI computing power in the future.
However, despite its early layout, many manufacturers have released MCUs with integrated AI computing power before the STM32N6 series. For example, Texas Instruments (TI) released the TMS320F28P55x controller of the C2000 series, the industry's first real-time MCU product with NPU. In addition, NXP has also launched the RT700, and i.MX 94 series in the past two months. Together with Infineon's PSoC Edge series. The competition in the MCU market is already extremely fierce even though the applications have not yet been fully popularized.
Many MCU manufacturers have their core technologies. Although ST's new products are full of sincerity, it is not easy to stand out in such fierce market competition. The "involution" of the Microcontroller market is no longer limited to basic products. The end-side AI market is also highly competitive. There are not only international manufacturers, but local manufacturers are also actively deploying. As a veteran in the MCU industry, ST's advantage lies in the deep accumulation of product lines and the leading position of ecological operations. The multiple sub-series derived from the STM32 series can already cover all-scenario applications. Let's overview the STM32 series product lines and their naming rules.
STM32 Product Lines Overview
STM32F Series: The best choice for high-performance and multi-function applications
The STM32F series is the most widely used series of STM32 products, mainly based on ARM Cortex-M0/M3/M4/M7 cores. This series provides a wealth of peripherals and interfaces, widely used in high-performance and multi-function scenarios such as industrial automation, consumer electronics, and medical equipment.
STM32F0xx: Low cost, low power consumption, designed for cost-sensitive and low-power applications.
STM32F1xx: Mid-to-low-end microcontrollers with rich peripherals and excellent performance.
STM32F2xx/F4xx/F7xx/F8xx: High-performance microcontrollers equipped with high-speed processor cores and rich peripherals, suitable for applications with high-performance requirements. Among them, the STM32F4 sub-series is based on the Cortex-M4 core, with a floating-point unit (FPU) and digital signal processing (DSP) instruction set; the STM32F7 sub-series is based on the Cortex-M7 core.
STM32G series: Ideal choice for cost-effective and multi-functional applications
The STM32G series is a general-purpose series in the STM32 product line, mainly based on the ARM Cortex-M4 and M0+ cores. This series provides a wealth of peripherals and interfaces, especially suitable for embedded applications with high cost-effectiveness and multi-functional requirements.
STM32G0xx: Based on the Cortex-M0+ core, it provides low cost and low power consumption, suitable for cost-sensitive applications.
STM32G4xx: Based on the Cortex-M4 core, it has high performance and rich peripherals, suitable for industrial control, motor control, consumer electronics, and other fields.
STM32L series: The best choice for low-power applications
The STM32L series features low power consumption, mainly based on the ARM Cortex-M0+/M3/M4 cores. The series has a variety of low-power modes, and the appropriate power consumption mode can be selected according to application requirements. It is suitable for applications that require long-term battery power.
STM32L0xx: Based on the Cortex-M0+ core, it further reduces power consumption and is suitable for ultra-low-power IoT devices.
STM32L1xx: Low-power microcontroller with rich peripherals and low-power features.
STM32L4xx: Based on the Cortex-M4 core, it combines high performance and low power consumption, and is suitable for applications that require long-term battery power, such as wearable devices, sensor nodes, and smart homes.
STM32H series: The best choice for high-performance applications
The STM32H series is a high-performance series in the STM32 product line, mainly based on the ARM Cortex-M7 core. They provide the highest performance and the richest peripherals. It is suitable for applications that require extremely high computing performance and rich peripherals. Among them, products such as STM32H7xx and H5xx are included.
STM32W series: MCU with a wireless communication function
The STM32W series is an MCU with a wireless communication function. It is based on the ARM Cortex-M core and integrates wireless communication modules such as Wi-Fi and Bluetooth. This series is suitable for applications that require wireless communication functions.
STM32WBxx: Integrated Wi-Fi and Bluetooth functions, based on the Cortex-M4 and M0+ dual-core architecture, supports Bluetooth and IEEE802.15.4 (such as Zigbee and Thread) communication protocols, suitable for smart home, wearable devices, industrial Internet of Things and other applications. Among them, sub-series such as STM32WB5 are included.
STM32WBA54/55 series: Wireless MCU, based on the Arm Cortex-M33 core, running at 100MHz, supporting the latest BLE5.4 and Matter standards, as well as 2.4GHz protocols such as Thread and ZigBee, and Bluetooth LE Audio low-power Bluetooth audio specifications.
STM32MP Microprocessor (MPU)
The STM32MP series of microprocessors integrates high-performance application processors and low-power real-time cores, suitable for industrial applications that require strong robustness in connected and secure environments.
STM32MP2 series: 64-bit MPU, integrated 1.35TOPS NPU, using 1/2 core 1.5GHz Cortex-A35 and a Cortex-M33 real-time core, and integrated with VeriSilicon's GC8000UL NPU/GPU. This series can flexibly run AI algorithms on CPU, GPU, or NPU, and is suitable for industries related to edge AI such as smart industry, smart home, smart gateway, smart city, smart transportation, etc.
In addition, they also launched ultra-low power entry-level MCUs such as STM32U0, as well as cost-effective MCUs of the STM32H7R/S series.
STM32 product lines cover various needs from low power consumption to high performance, from simple applications to complex systems. Developers can flexibly choose the appropriate model and series according to the specific applications and performance requirements.