ARM adds custom instructions to M85 controller in RISC-V AI fightback
Technology News | March 13, 2023
By Nick Flaherty
RISC-V IOT AIMPUS/MCUS
ARM has developed a new version of its high end Cortex-M85 microcontroller, allowing chip designers to add custom instructions for AI applications.
This has been one of the key advantages of the competing RISC-V instruction set and signals a fight back by ARM ahead of the Embedded World (EW2023) exhibition in Nuremberg, Germany, this week.
Renesas Electronics will be showing the first live demonstrations of artificial intelligence (AI) and machine learning (ML) implementations on the previous generation M85 design.
This first demonstration showcases a people detection application developed in collaboration with UK ML startup Plumerai that identifies and tracks persons in the camera frame in varying lighting and environmental conditions.
The second demo showcases a motor control predictive maintenance use case with an AI-based unbalanced load detection application using Tensorflow Lite for Microcontrollers with CMSIS-NN.
The latest version of the M85 core, Cortex-M85 r1, includes ARM Custom Instructions (ACI) which allow designers to include custom defined data processing instructions directly on the controller. These were first proposed in 2019 in the ARMv8.1 architecture but are only now being implemented in designs, in part to address the increasing popularity of cores based on the customisable RISC-V instruction set.
“With ACI on Cortex-M85, the user does not have to turn to alternative architectures to implement a desired instruction encoding. Instead, this can now be done on CPUs that are based on the ARM architecture, with Cortex-M85 being the first high-performing microcontroller to provide this option.,” said ARM. “Through ACI, the user is given the power to innovate within the proven AEM architecture, while maintaining the ecosystem advantages of the Cortex-M CPUs.”
ACI is suitable for applications using specialized bit field processing, trigonometric functions, and image pixel manipulations and can also be applied to accelerate frequently used data processing functions. It is also being implemented on the Cortecx-M55 core.
AI demonstrations
In the meantime, Renesas is taking advantage of the Helium technology in the core which will be part of the RA (Renesas Advanced) family of MCUs.
Helium is ARM’s M-Profile Vector Extension, available as part of the Armv8.1M architecture and provides a boost for machine learning (ML) and digital signal processing (DSP) applications, accelerating applications such as endpoint AI.
“We’re proud to again lead the industry in implementing the powerful new Arm Cortex-M85 processor with Helium technology,” said Roger Wendelken, Senior Vice President in Renesas’ IoT and Infrastructure Business Unit.
“By showcasing the performance of AI on the new processor, we are highlighting technical advantages of the new platform and at the same time demonstrating Renesas’ strengths in providing solutions for emerging applications with our innovative ecosystem partners.”
“We’re excited to take part in this ground-breaking demonstration,” said Roeland Nusselder, CEO of Plumerai. “ARM’s Helium technology supported on the new RA MCUs with the Cortex-M85 core significantly accelerates the Plumerai inference engine.”
“This performance uplift will enable our customers to use larger and more accurate versions of Plumerai’s People Detection AI, add additional product features, and extend battery life,” he said. ”Our customers have an insatiable appetite for adding new and more accurate AI features that run on a microcontroller. Together with Renesas, we are the first to fulfil this demand.”
The Cortex-M85 core supports ARM TrustZone alongside a Renesas integrated cryptographic engine, immutable storage, key management and tamper protection against DPA/SPA side-channel attacks. The Armv8-M architecture also brings Pointer Authentication/Branch Target Identification (PAC/BTI) security extension, a new architectural feature that provides enhanced mitigation from software attack threats and helps achieve PSA Certified Level 2 certification.
The new RA MCUs based on the Cortex-M85 core will be supported by Renesas’ Flexible Software Package (FSP). This enables faster application development by providing all the infrastructure software needed, including multiple RTOS, BSP, peripheral drivers, middleware, connectivity, networking, and security stacks as well as reference software to build complex AI, motor control and graphics solutions.
It also allows developers to integrate their own legacy code and choice of RTOS with FSP, providing full flexibility in application development.