In October 2021, Arm launched Arm Virtual Hardware (AVH), a cloud-based offering which enables software development without the need for physical hardware. This helps reduce time to market for embedded software developers and simplifying end-device integration into IoT services.
A little over a year ago, we extended the capabilities of AVH to address new uses cases and to enable a wider range of Arm processors and third-party hardware via Corellium’s hypervisor technology. This included adding hardware from partners NXP Semiconductors, STMicroelectronics and Raspberry Pi, as well as Arm models of Corstone-300, Corstone-310, and Cortex-M processors ranging from Cortex-M0 to Cortex-M33. Over the past year, hundreds of embedded and IoT developers across the Arm ecosystem have participated in a private beta with this powerful new AVH offering, incorporating it into their development workflows, CI/CD pipelines, IoT SaaS solutions, and development tools. Our private beta users have also provided invaluable feedback that has helped improve and enhance the AVH service.
Today we are pleased to announce that this service has transitioned from private beta to public beta and is now open to anyone with an Arm account to try out and use for commercial purposes. The public beta is available for a trial period of 30 days followed by a paid service based on usage per device-hour. Go to arm.com/virtual-hardware today to get started.
As we expand access to AVH, it is also worth reflecting on some of the new, enhanced software development capabilities we have enabled over the past year. We recently added the NXP i.MX 93 Applications Processor to the AVH portfolio. This enables faster and earlier software development for developers using this board.
NXP i.MX 93 Applications Processor
A further enhancement of AVH is the addition of Arm Cortex-M and Corstone AVH models, which are available as a service through the SaaS platform, making the complete portfolio accessible with web console or API.
Board and models available through the SaaS platform
This uniform access to all types of AVH removes dependencies on local or cloud computing infrastructure to run AVH instances.
Beyond enhancing the AVH portfolio, we have built on a solid foundation for AVH integration into CI/CD, IoT, and MLOps services.
Following the announcement of our partnership with GitHub last year, we have now jointly enabled a native GitHub Actions Runner that includes Arm Cortex-M and Corstone Fixed Virtual Platforms (FVP) with AVH, and the Arm Compiler to allow developers to seamlessly create Arm CI/CD pipelines in their GitHub environment. This integration is available today as private beta for GitHub Enterprise customers. Request access here.
We have also enabled sound CI/CD practices for foundational IoT software with AVH. This is best exemplified with GitHub Actions-based flows for:
Learn more about what you can do with Matter and AVH in this blog.
Our AVH story, which started with AWS, has truly flourished and continues to blossom with:
How to build, train, test, and deploy machine learning applications on AVH.
We look forward to even more capabilities with AVH and AWS in the coming year.
Earlier this year we launched an integration with Remote.It, leveraging its cutting-edge IoT services. Together, AVH and Remote.It – a simple solution to remotely access your virtual Arm devices hosted in AVH – are powerful tools that can help developers streamline their workflows and bring products to market faster than ever.
Arm Tech Talk from Remote.it: Simple and secure remote access to Arm Virtual Hardware devices
Through close partnership with Arm, TDK Qeexo is working to extend our AVH integration, adding support for additional Arm Corstone platforms, and incorporating Arm's Synchronous Data Streaming (SDS) framework. This enables full machine learning training and validation workflows using existing sensor data, executed across a variety of Arm processors.
Over the last year, Arm has deepened its collaboration with Baidu PaddlePaddle to accelerate edge AI development and deployment on Arm. We not only showcased how to deploy PaddlePaddle AI framework and models, including optical character recognition (OCR) on Arm Cortex-M55 through Arm Corstone-300 AVH, but also sponsored the Paddle Hackathon 2023 to enlist the PaddlePaddle community to help expand the model zoo for edge AI on Arm. We were pleasantly surprised to receive the first submission for code review within a few hours after the hackathon projects were posted, thanks to the agility brought by AVH. Previously it would have taken hours and maybe even weeks to buy the hardware development board and set it up properly to get started.
Stay tuned, AVH will soon be hosted natively on Baidu Cloud, and we cannot wait for edge AI developers in China to check that new offering out.
Arm Tech Talk from Baidu: How to apply AI to OCR text recognition
As you can see, we have had an exciting year working with developers and partners to broadly adopt AVH into their development workflows, services, and solutions. We cannot wait to see what developers will do with even broader access to AVH whether through our public beta, the GitHub private beta, AWS, or several other ecosystem partner integrations.
Register for AVH public beta