Dual Arm Cortex-M4, RISC-V processor integration
CNN accelerator, 1/100th the power consumption of MCU


Mouser Electronics announced on the 11th that it is supplying the 'MAX78000' SoC, which integrates Maxim Integrated's dual-core MCU and an ultra-low-power deep neural network accelerator.

The MAX78000 provides the computing power required for high-performance AI applications. It is ideal for applications such as machine vision, facial recognition, object detection and classification, time-series data processing, and audio processing.

Two Arm Cortex-M4 processors and a 32-bit RISC-V coprocessor were integrated for system control. In particular, the RISC-V processor supports special features that enable high-speed and low-power data loading for neural network accelerators. Once data is configured and loaded, the 442KB Convolutional Neural Network (CNN) accelerator executes AI inference at less than 1/100th of the power consumption and more than 100 times faster speed compared to an MCU.

In addition, it maximizes the operating time of battery-powered IoT devices through high-efficiency power management features. Through dynamic voltage scaling, the processor minimizes active core power consumption, and selectable SRAM retention is available in low-power mode. This combination of low-power performance and high-efficiency computing reduces latency by up to 100-fold and enables AI inference to be executed at the IoT edge.

Mouser also supports the MAX78000 evaluation kit. In addition to the MAX78000 SoC, the evaluation kit features a digital microphone, gyroscope and accelerometer, support for parallel camera modules, and a 3.5-inch touch-sensitive color TFT display. A MAX78000FTHR development board in the Adafruit Feather form factor, designed for rapid prototyping of small systems, is also available.