Author: Jean-Luc Aufranc (CNXSoft)

Now even endpoints like sensors are capable of running basic artificial intelligence workloads thanks to microcontroller-class chips with built-in AI accelerators or instructions such as Kneron KL720 Arm Cortex-M4 AI SoC, GAP8 RISC-V IoT processor, or Cortex-M55 MCU core coupled with Ethos U55 MicroNPU. Maxim Integrated has now launched its own AI capable microcontroller with MAX78000 combining a Cortex-M4F core, a 32-bit RISC-V core, and a CNN accelerator enabling AI inferences at less than 1/100th the energy or 100 times the speed of software solutions making especially suited for battery-powered AI applications. MAX78000 specifications: Ultra-Low-Power microcontrollers Arm Cortex-M4F core with FPU Up to 100MHz with 512KB Flash and 128KB SRAM 32-Bit RISC-V ultra-low-power co-processor up to 60MHz Neural Network Accelerator optimized for Deep Convolutional Neural Networks​​ (CNN) 442k 8bit Weight Capacity with 1,2,4,8-bit Weights Input Image Size up to 1024 x 1024 pixels Network Depth up to 64 Layers Layer […]