Arm announces high-performing internet of things (IoT) hardware

Earlier this week, Arm debuted several new products for its internet of things (IoT) portfolio, including its highest-performing Cortex-M microcontroller yet. The new updates span Arm’s Total Solutions for IoT roadmap and target applications such as cloud-native edge devices and voice recognition. Overall, Arm’s goal is to support the IoT ecosystem, which ranges from sensors to industrial applications.

Total Solutions for IoT

Arm launched its Total Solutions for IoT half a year ago with the plans to deliver a full-stack solution to accelerate IoT development. The platform combines IP, software, machine learning (ML) and other tools for product design. As is common in Arm’s business model, the company is focused on providing common hardware IP that allows developers to innovate in areas of differentiation. Here, the core of its IoT platform is Arm Corstone as a pre-integrated and pre-verified IP subsystem.

Arm is now launching two new Total Solutions. The first, Total Solution for Cloud Native Edge Devices, is based on Arm’s most high-end Corstone-1000 platform, which features Cortex-A processors such as the Cortex-A53, but also leverages Cortex-M for the most efficient performance. The company  says the addition of the Cortex-A32 provides IoT developers the capability of running operating systems like Linux as well as application-class workloads. Potential devices that could benefit from the newly introduced tech solution include wearables, gateways and smart cameras. In addition, the platform also contains a hardware secure enclave for the security of sensitive data.

The company’s second release, the Total Solution for Voice Recognition, is based on the Corstone-310 subsystem, which contains the new Cortex-M85 and the Ethos-U55 NPU to deliver what Arm claims is its highest performance MCU-based design yet. This solution could be used to enhance technologies like smart speakers, thermostats, drones and even factory robots.

The new Cortex-M85 is Arm’s highest-performing Cortex-M processor to date, with a claimed uplift of 30% over the Cortex-M7 and 20% for ML workloads. It supports the Armv8.1-M instruction set that features Arm Helium technology for endpoint ML and digital signal processing (DSP) workloads. With Helium, performance is improved anywhere from five  to 15 times with the introduction of new, low precision scalar and vector instructions. 

The Cortex-M85 also features Arm TrustZone support for enhanced security. Arm says it achieves the PSA Certified Level 2 security baseline for IoT deployments.

“The IoT runs on Arm and we have a responsibility to create greater opportunities for IoT innovation and scale by continually raising the bar on performance, simplified development and software reuse for our ecosystem,” said Mohamed Awad, vice president of IoT and embedded technology  at Arm. 

Virtual hardware

This week, Arm also revealed several new virtual devices to further expand its virtual hardware. The additions include the Corstone platforms and seven new Cortex-M processors, up to the Cortex-M33. Arm also reports that it’s expanding its library with third-party hardware from partners such as NXP, ST Microelectronics and Raspberry Pi. 

Arm Virtual Hardware, which launched together with its Total Solutions last fall, enables software development already in advance of silicon through a cloud-based offering. This allows the Arm ecosystem to adopt cloud-based development. 

Open IoT

Arm also launched Project Centauri to foster standardized IoT development. To that end, Arm has announced that it’s delivering the first release of the Open IoT SDK Framework that contains the new Open-CMSIS-CDI software standard. This standard defines a Common Device Interface (CDI) for the Cortex-M ecosystem. Arm says that eight industry players are already involved, with cloud service providers, ODMs and OEMs among those.

All of the company’s new tech solutions are immediately available for licensing and can be accessed in the cloud. As part of Arm’s roadmap, the company is also working on Total Solutions for vision, object recognition and smart sensor fusion. The first one will be addressed by Cortex-A53, while the latter two will leverage Cortex-M processors.