- Arm Cortex-M55 boosts ML performance by up to a 15x and DSP performance by up to 5x
- Arm Ethos-U55 in combination with Cortex-M55 can increase ML performance by up to 480x
Arm, an enabler of intelligent computing and IoT technology, has made two new additions to its artificial intelligence (AI) platform: the Arm Cortex-M55 processor, a new machine learning (ML) IP and Arm Ethos-U55 NPU, a micro NPU (Neural Processing Unit) for Cortex-M, designed to deliver a 480x ML performance to microcontrollers.
As IoT and AI are increasingly merging, along with the rollout of 5G, more on-device intelligence means that smaller, cost-sensitive devices that are smarter and more capable can provide greater privacy and reliability due to less reliance on the cloud or internet. By delivering this intelligence to microcontrollers, silicon and development costs can be reduced and time to market for product manufacturers can be increased for efficiently enhancing on-device digital signal processing (DSP) and ML capabilities.
Arm’s new IP and supporting toolchain encompasses remarkable levels of on-device ML processing for billions of small, power-constrained IoT and embedded devices. This helps in empowering AI hardware and software developers with more new ways to innovate.
“Enabling AI everywhere requires device makers and developers to deliver machine learning locally on billions, and ultimately trillions of devices,” said Dipti Vachani, senior vice president and general manager, Automotive and IoT Line of Business, Arm. “With these additions to our AI platform, no device is left behind as on-device ML on the tiniest devices will be the new normal, unleashing the potential of AI securely across a vast range of life-changing applications.”
AI-capable Cortex-M processor
Cortex-M55 offers to be Arm’s most AI-capable Cortex-M processor ever and the first based on the Armv8.1-M architecture with Arm Helium vector processing technology for enhanced, energy-efficient DSP and ML performance.
Cortex-M55 boosts ML performance by up to a 15x and DSP performance by up to 5x with better efficiency, as compared to previous Cortex-M generations.
Additionally, Arm Custom Instructions, a new feature of Cortex-M processors, can be implemented to extend processor capabilities for specific workload optimisation.
Arm’s first micro NPU for Cortex-M
The Ethos-U55 is highly configurable and specifically designed for accelerating ML-based features in space-constrained embedded and IoT devices. Its advanced compression techniques save power and reduce ML model sizes significantly to enable execution of neural networks that previously ran only on larger systems.
For even more demanding ML systems, the Cortex-M55 can be easily paired with the Ethos-U55 to deliver a combined 480x increase in ML performance.
Securing ML workloads
Cortex-M55 and Ethos-U55 are fully supported by Arm’s Cortex-M software toolchain. This ensures a unified development flow for traditional DSP and ML workloads. At the same time, specific integration and optimisations for leading ML frameworks, such as TensorFlow Lite Micro, will ensure that developers have a seamless experience and get the best performance from any Cortex-M and Ethos-U55 configuration.
Arm considers security to be the first-most priority for IoT to scale and not as an afterthought. Therefore, these processors and the accompanying Corstone reference design work with Arm TrustZone to ensure that security can be more easily incorporated into the complete system-on-chip.
The new Cortex-M CPU and the Ethos micro NPU highlight Arm’s commitment to empower silicon designers and device manufacturers with the agility to innovate on the Arm architecture, at even the smallest endpoints, and as a result, initiate a new wave of IoT creativity and innovation.