Low-Power Microcontroller Subsystems with ML Acceleration
The addition of Arm Helium vector processing technology in Arm Cortex-M55 and Cortex-M85 allows academics to run more demanding, yet energy-efficient, digital signal processing (DSP) and machine learning (ML) experiments. When coupled with the Arm Ethos-U55 accelerator, research into voice activation, image classification, and gesture recognition becomes accessible. Arm Fixed Virtual Platforms, affordable hardware, and a comprehensive toolchain enable seamless software development, as common signal processing functions are included in the libraries.
Hardware Design
Academics can access commercially proven, unencrypted, ML-oriented Cortex-M RTL, and pre-verified Corstone reference packages to reduce the cost and complexity of integrating custom research IP. Community proven reference designs, along with training, education, and support help streamline and accelerate the hardware journey.
Featured Resources
IP | Arm Academic Access
Institutional Access to a Suite of ML-orientated Cortex-M Technology
Arm Academic Access is a free subscription program, allowing multiple research teams to work with a wide range of technology, including Cortex-M55, M85, Ethos-U55, and U65 and associated Corstone packages.
COMMUNITY | SoC Labs
Leverage an ML-Orientated Cortex-M Reference Design
Based on the Corstone-300 reference package, this academic community platform was designed to support Taiwanese universities and is now available for other global research groups to explore.
ML Developers Guide to Cortex-M and Ethos-U
This guide provides an overview of ML development concepts and processes, details of software and hardware design considerations, links to tools and models, and related resources on analyzing and optimizing performance.
Community Resources
WEBINAR | Featured Webinars
Access an Academic SoC Webinar
Arm and CMC Microsystems will discuss how you can utilize our collective offers to enable your research work.
COMMUNITY | SoC Labs
Access Training on an ML-Orientated Cortex-M Reference Design
Express interest in training on the ML-orientated Cortex-M academic community platform.
COMMUNITY | SoC Labs
Community Support for Your SoC Project
Discuss your research with an experienced team at SoC Labs, the global academic community for Arm-based projects. Sign up to start the conversation.
Software Development
The ecosystem for developing software applications that support the Arm Helium vector extensions and/or Ethos NPU acceleration is rich and deep. The toolchain is available through a free Keil MDK Visual Studio Code extension or as standalone components like CMSIS and Arm Development Studio. Functionally accurate Arm Fixed Virtual Platforms are available for all Cortex-M processors and, as an academic, you can join the Developer Program for additional support.
Arm Tools and Developer Resources
TOOL | Fixed Virtual Platforms
Access Cortex-M Fixed Virtual Platforms
Academic developers can test and run software quickly and easily without the need for an Ethos-capable development board using Corstone-300 functionally accurate Fixed Virtual Platforms.
DEVELOPER | Arm Developer
Get Institution-Wide Access to Arm Development Studio
Access a dedicated IDE including the Arm Compiler for embedded and optimized libraries across Cortex-M applications. Universities can obtain a free institution-wide license.
DEVELOPER | Arm Developer
Join the Arm Developer Program
Join the free Arm Developer Program for community support from Arm and industry experts.
Featured Resources
TOOL | Arm Keil MDK
Cortex-M Toolchain
Access a comprehensive development environment of APIs, software components, debugging tools, and workflows for Cortex-M developers as Visual Studio Code extensions. Ideal for academic developers who want to get started in minutes.
Source Code and Instructions for Running ML Examples
The Arm Machine Learning Evaluation Kit contains several example applications, such as keyword spotting and object detection, that can be run performantly on Arm Cortex-M microprocessors with Helium vector extensions. Linux is the preferred environment to run these examples with limited Windows support.
BOOK | Featured Book
Fundamentals of M-Profile Vector Extensions E-Book
Covering the fundamentals of SIMD, this free e-textbook provides a programmers' view of the Cortex-M profile vector extensions, including guides on debugging, optimizing and running ML workloads.
Other Resources
RESOURCES | Arm Model Zoo
ML Models Optimized for Cortex-M
A comprehensive list of free and accessible ML models that will run on Cortex-M microprocessors with Helium vector extensions and Ethos-U NPUs, as well as other Arm platforms.
RESOURCES | Learning Paths
Deploy Yolo v8 on a Fixed Virtual Platform
Learn how to get the computer vision model Yolo v8 up and running on a Corstone-300 Fixed Virtual Platform, so that you can develop objection detection, classification, and segmentation software applications.
RESOURCES | Learning Paths
ML Development with Ethos-U Processors
Learn how to select physical and virtual hardware targets for ML application development with Cortex-M and Ethos-U technology, install software tools, and identify and draw insights from existing example applications.
Our comprehensive range of online courses, books, and education kits is useful not only for teachers and learners, but for researchers as well.
Other academic developer teaching and research platforms.
Education content for teachers and researchers.
Education content for undergraduate and early career learners.
Training for those who already have access to Arm technology.