To keep the battery from running down, Google's new Android 4.4 KitKat operating system mandates that smart sensors provide real-time contextual awareness functions while the application processor is off. (Source: Google)
Yeah. Apple looks very different. Its M7 used NXP's Cortex-M3 based LPC18A1, of which the ARM core is optimized for performance efficiency. I guess the teams from both Apple and NXP have done a lot of things to improve the coprocessor's power efficiency.
It is exciting that most MCU/MEMS vendors have been showing their muscles since Apple released its latest iPad product with M7 sensor hub coprocessor. It seems most of the exisiting solutions are ARM Cortex-M0/0+ based or comparable solutions, which target extemely energy efficiency but with limited computing capabilities.
Will Cortex-M3/4 based solutions be coming out in the future to meet ever-increasing more complex computing tasks, e.g., 3-D recognition, while still tailored for low energy requirement by other advanced low power design techniques?
This seems like a very incremental announcement of a so-so drop-in improvement part, in Q4 Invensense is due to have a 9 axis drop-in part which I would far prefer. MEMs inertial parts drift badly and without the magnetometer which the new 6 axis MPU-6515 doesn't have, it is going to get lost rather quickly. If you need contextual awareness in a smart device then a 9 axis part in Q1 will have a far better solution.