The Arm Cortex-M55 processor and Arm Ethos-U55 NPU (Neural Processing Unit) for the Cortex-M range can deliver a combined 480x leap in ML performance.
By increasing intelligence on MCUs, Arm is trying to reduce silicon and development costs and speed up time to market for designs using DSP and ML technology.
The Cortex-M55 is based on the Armv8.1-M architecture with Arm Helium vector processing technology for enhanced, energy-efficient DSP and ML performance. The Cortex-M55 architecture provides up to a 15x better ML performance and a 5x better DSP performance over previous Cortex-M generations, while still being more power-efficient.
Arm Custom Instructions are also available to extend processor capabilities for specific workload optimization.
For higher ML system performance, the Cortex-M55 can be paired with the Ethos-U55, Arm’s first microNPU. Together the two IPs can provide a combined 480x increase in ML performance over existing Cortex-M processors.
Ethos-U55 is configurable and has been specifically developed to accelerate ML inference in area-constrained embedded and IoT devices. The processor offers advanced compression techniques to save power and reduce ML model sizes, while allowing execution of neural networks that previously only ran on larger systems.
Cortex-M55 and Ethos-U55 are supported by Arm’s Cortex-M software toolchain to ensure a unified development flow for traditional, DSP and ML workloads. Specific integration and optimizations for leading machine learning frameworks, starting with TensorFlow Lite Micro, will also ensure that developers have a seamless experience and get the best performance from any Cortex-M and Ethos-U55 configuration.
These processors and the accompanying Corstone reference design work with Arm TrustZone to ensure security can be more easily incorporated into the final system-on-chip.