Single processor solutions will inevitably have to trade-off the processor's computational power against its efficiency. When that trade-off becomes too great, developers look towards dual-core solutions.
Tomorrow’s commercial products will demand connectivity solutions to support advanced, highly-complex applications while also reducing time-to-market and development risk.
Two cores better than one
In a dual-core wireless SoC, one powerful processor with a high clock rate can be dedicated to the computational heavy lifting (and then switched off to save power until it is next needed). The other, a power-optimized processor with a slower clock, can be dedicated to wireless networking.
Better yet, each core can also run its own software, preventing clashes between application code and RF protocol and allowing optimized algorithms for their destined processor. To make things even more battery-friendly, each processor can benefit from its own power management system.
Moreover, the application code can run continuously without interruptions from the radio. Or the other way around, the radio can run without being interrupted by the application code. Such a configuration is useful in mesh (for example, lighting) and mission-critical motor control applications.
An SoC for today … and tomorrow
Nordic’s nRF5340 is the world’s first wireless SoC to integrate two user-accessible Arm Cortex-M33 processor cores. The application processor is optimized for performance and can run at either 128 (for 514 CoreMark performance) or 64 MHz, and the network processor is optimized for efficiency (101 CoreMark/mA at 64 MHz).
The nRF5340’ dual-core architecture has been cleverly divided into distinctly defined subsystems:
- An efficient and stable subsystem based on the network processor and used to supervise the RF protocol software (for example, a Bluetooth LE stack)
- The application processor subsystem, with th