Close
Close

Published

ARM and XMOS push down cost and power of end-point machine learning

A hot area in chip design is currently tinyML – the quest to squeeze machine learning and other AI capabilities into the smallest devices, to add intelligence into end-points while keeping cost and power consumption low. In the past week, ARM and UK-based XMOS have both announced new developments in this area, targeting the Internet of Things (IoT), and joining Qualcomm and others in a crowded field. ARM has unveiled its two latest IP cores with end-point machine learning (ML) in mind. One is the new addition to its Cortex-M microcontroller core family, the M55, while that is paired with an accelerator called Ethos-U55, which is a miniature version of the company’s Ethos neural processing unit (NPU) range. The two…

Close