Ultra-Low-Power MAX78000 Arm® Cortex®-M4 Processor with Neural Network Accelerator

Artificial intelligence (AI) requires extreme computational horsepower, which comes at a cost in terms of power consumption and heat. ADI is cutting the power cord for AI insights. The MAX78000 is a new breed of AI microcontroller built to enable neural networks to execute at ultra-low power and live at the edge of the IoT. As applications demand more intelligence at the edge, this product combines the most energy-efficient AI processing with ADI's proven ultra-low-power microcontrollers. ADI's hardware-based convolutional neural network (CNN) accelerator enables battery-powered applications to execute AI inferences while spending only microjoules of energy.

ADI's MAX78000 is an advanced system-on-chip featuring an Arm Cortex-M4 with FPU CPU for efficient system control with an ultra-low-power deep neural network accelerator. The CNN engine has a weight storage memory of 442 KB and can support 1-, 2-, 4-, and 8-bit weights (supporting networks of up to 3.5 million weights). The CNN weight memory is SRAM-based, so AI network updates can be made on the fly. The CNN architecture is highly flexible, allowing networks to be trained in conventional toolsets like PyTorch and TensorFlow®, then converted for execution on the MAX78000 using tools provided by ADI.

In addition to the CNN engine memory, the MAX78000 has large on-chip system memory for the microcontroller core, with 512 KB Flash and up to 128 KB SRAM. Multiple high-speed and low-power communications interfaces are supported, including I²S and a parallel camera interface (PCIF).

The device is available in 81-pin CTBGA (8 mm x 8 mm, 0.8 mm pitch) and 130-pin WLP (4.6 mm x 3.7 mm, 0.35 mm pitch) packages.