Arm has announced 'Arm Helium technology' that intends to deliver enhanced machine learning and signal processing for the smallest embedded devices.

The M-Profile Vector Extension (MVE) for the Arm Cortex-M series processors have been designed to enhance the compute performance of the Armv8.1-M architecture on the secure foundation of Arm TrustZone.

According to Arm, Helium will deliver up to 15x more ML performance and up to 5x uplift to signal processing for future Arm Cortex-M processors.

Advanced digital signal processing (DSP) is available through Arm Neon technology in richer Cortex-A based devices. For more constrained applications, Arm has also added DSP extensions in its higher performance Cortex-M processors (Cortex-M4, Cortex-M7, Cortex-M33 and Cortex-M35P). Both technologies can be used to accelerate ML compute in certain applications.

For the most constrained embedded systems where energy efficiency is prioritised, historically the solution has been coupling a Cortex processor with a DSP in SoCs, which adds complexity to both hardware and software design.

Armv8.1-M with Helium is designed to eliminate these challenges by delivering real-time control code, ML and DSP execution without compromising efficiency.

The idea is that it will provide software developers with the ability to securely scale intelligent applications that take advantage of DSP capabilities across a wider range of devices. This should enable enhanced support for emerging applications across three key categories; vibration and motion, voice and sound, and vision and image processing.

The hope is that this will improve the user experience in future devices such as sensor hubs, wearables, audio devices and industrial applications powered by next-generation SoCs based on Cortex-M with Helium technology.

Software development will be made simpler due to Helium’s unified tool chain, libraries and models. The Helium toolchain includes the Arm Development Studio, encompassing Arm Keil MDK, Arm Models (which are immediately available to developers for code modelling) and various software libraries including CMSIS-DSP and CMSIS-NN.

For signal processing applications Arm has simplified the process, by removing the need for a dedicated DSP or function accelerator and eliminating another layer of design complexity.

Audio Analytic, a partner of Arm, had early access to the new extensions. According to the company, its sound recognition software (ai3) will now be at least 50% faster when running on chips based on the new Armv8.1-M architecture.

Commenting on the launch, Dr Chris Mitchell, Audio Analytic’s CEO and founder, said: “There is considerable demand to run advanced AI, like sound recognition, at the edge. Principally because cloud infrastructure is expensive and edge-based processing offers privacy benefits for end users. Now, thanks to Arm, consumer and IoT devices can deliver supercharged AI at even lower-power and lower-cost. The net result is being able to fit more features onto a device or being able to offer AI on an AA battery.”