NXP’s cross-industry MCU products are positioned as cross-border SoCs that combine the low cost of MCUs with the high performance of MPUs. Let’s explore the architecture of cross-industry MCUs with the guidance of Jingxin.
1. i.MX RT700 Series
The i.MX RT700 series features an Arm Cortex-M33 core with a clock speed of 325MHz, HiFi 4 DSP, eIQ Neutron NPU (Neural Processing Unit), and a dedicated perception computing subsystem.
With the integrated eIQ Neutron NPU, AI inference speed is improved by 172 times compared to pure CPU, and power consumption is reduced by 119 times, supporting multimodal AI (such as presence detection and gesture recognition).
Utilizing adaptive deep sleep technology, power consumption is reduced by 30%-70% compared to previous generations, and it integrates 7.5MB of ultra-low power SRAM, making it suitable for battery-powered smart wearables, medical devices, smart home HMIs, voice interaction devices, and portable health monitoring terminals.
2. i.MX RT600 Series combines the Cortex-M33 core with HiFi 4 DSP, optimized for audio and machine learning, supporting 4.5MB of on-chip SRAM.
It is compatible with the NXP eIQ machine learning compiler and integrates a library of voice preprocessing algorithms (such as Alango and Sensory). Power management is optimized, making it suitable for edge voice recognition and sensor fusion scenarios.
3. i.MX RT500 Series integrates the Cortex-M33 core with Tensilica Fusion F1 DSP, supporting 5MB of on-chip SRAM. It features a built-in 2D GPU, suitable for lightweight graphics rendering needs.
It supports dynamic voltage frequency scaling (DVFS) to extend battery life, making it suitable for smart thermostats, portable medical devices, and low-power IoT gateways.