Dual-core MCUs with Arm Cortex-M85, M33, and Ethos-U55 NPU deliver 7300+ CoreMarks, MRAM-based speed and endurance, AI-optimized peripherals, and a new software framework for seamless AI deployment.
Renesas Electronics has unveiled the RA8P1 MCU Group, its most powerful microcontroller line yet, built for high-performance edge AI, machine learning, and real-time analytics applications. Featuring a combination of 1GHz Arm Cortex-M85, 250MHz Cortex-M33, and the Ethos-U55 NPU, the RA8P1 delivers over 7300 CoreMarks and 256 GOPS AI performance—setting a new benchmark in the MCU world.
Optimized for CNNs and RNNs, the Ethos-U55 NPU offloads compute-intensive AI tasks from the CPU, delivering up to 35x faster inference speeds depending on the model. Compatible with popular networks like ResNet, DS-CNN, MobileNet, and TinyYolo, it supports complex voice and vision applications at the edge.The key features are:
- Parallel & MIPI-CSI2 camera interfaces for vision AI
- Secure boot with immutable first-stage bootloader
- Timers: 32/16-bit high-res, ultra-low-power 32-bit
- Analog: 16-bit ADC, 12-bit DAC, high-speed comparators
- Connectivity: Gigabit Ethernet + TSN switch, USB 2.0 HS/FS, CAN FD
- Serial interfaces: SDHI, SPI, I3C, I2C
- External memory: 32/16-bit interface (CS/SDRAM)
Built on TSMC’s 22ULL (ultra-low leakage) process, the RA8P1 integrates embedded MRAM for faster write speeds and greater endurance over Flash. It includes 2MB ECC-protected SRAM and 1MB MRAM, with SIP flash options up to 8MB for advanced AI workloads. A 16-bit CEU supports up to 5MP image sensors, while MIPI CSI-2 and rich audio interfaces power voice and vision use cases. Security is equally robust.
Applications include an AI Drawing Robot Arm that leverages precision motor control and smart image processing to enable creative and automated design tasks. In communication, an AI-powered video conferencing camera integrates an AI-accelerated MCU, wireless connectivity, high-fidelity audio, and flexible power options. Its 360° coverage with dual cameras and beamforming audio ensures immersive conferencing experiences. For smart appliances and industrial systems, an e-AI-ready GUI controller offers intelligent HMI capabilities, predictive maintenance, and extensive connectivity. Additionally, an AI-enabled surveillance camera powered by a high-performance MPU with DRP-AI delivers advanced 4K imaging, strong AI processing, and energy-efficient operation for enhanced security applications.
The company is also launching RUHMI, a unified framework supporting TensorFlow Lite, ONNX, and PyTorch for seamless AI deployment across MCUs and MPUs. RUHMI integrates with Renesas e² Studio IDE, simplifying model optimization and application development.
The MCUs are shipping now, alongside evaluation kits and software examples. Partner solutions from Nota.AI, Irida Labs, and others are also available. Explore more details by clicking here.
