Arm Holdings Limited (NASDAQ: ARM, hereinafter referred to as “Arm”) today announced the launch of the Arm® Ethos™-U85 Neural Processing Unit (NPU), which is Arm’s highest performance and most energy-efficient Ethos NPU product to date, along with a new IoT reference design platform—Arm Corstone™-320—to accelerate the deployment of voice, audio, and visual systems.
Paul Williamson, Senior Vice President and General Manager of Arm’s IoT Division, stated:“As the scale of edge artificial intelligence (AI) deployment continues to expand, chip innovators must address increasingly complex systems and software, the surging demand for AI performance, and the pressure to accelerate time-to-market. Meanwhile, software developers require a more consistent and streamlined development experience, with easier integration with new AI frameworks and libraries. The new technologies launched by Arm meet the growing demands for accelerating edge AI deployment. From startups to the world’s largest microcontroller manufacturers, every company views Arm technology as the preferred platform for building their AI solutions from cloud to edge.”
Image: Arm Empowers Continuous Innovation in Edge AI
Arm Ethos-U85 NPU:
Arm’s New Third-Generation NPU for Edge AI
Compared to the previous generation, the Ethos-U85 offers a fourfold performance increase and a 20% improvement in energy efficiency, with its MAC units expandable from 128 to 2048 (achieving 4 TOPs of computing power at 1GHz), providing robust support for applications requiring higher performance, such as factory automation and commercial or smart home cameras. The Ethos-U85 offers the same consistent toolchain, allowing partners to leverage existing investments for a seamless developer experience. More importantly, the new Ethos-U85 NPU supports AI frameworks such as TensorFlow Lite and PyTorch.
Image: The Arm Ethos-U85 NPU will provide the computing power needed to execute leading AI functions on edge and endpoint devices
Ethos-U85 supports Transformer architectures and Convolutional Neural Networks (CNN) for AI inference. The Transformer architecture will drive new applications, particularly effective in visual and generative AI use cases for understanding video, filling in missing parts of images, or analyzing data from multiple cameras for tasks such as image classification and object detection.
As microprocessors are deployed in more high-performance IoT systems such as industrial machine vision, wearables, and consumer robotics, the Ethos-U85 is designed to work in conjunction with Arm’s leading Armv9 Cortex®-A CPU to accelerate machine learning (ML) tasks and provide high-efficiency edge inference capabilities for a broader range of higher-performance devices.
To date, over 20 authorized licensing partners have adopted the Arm Ethos NPU series, with Alif Semiconductor and Infineon being early adopters of the new Arm Ethos-U85 NPU.
Reza Kazerounian, Co-Founder and President of Alif Semiconductor, stated:“The ML workloads of the new generation of edge AI applications require high-performance delivery in an energy-efficient manner. Alif is the first company in the industry to launch edge AI solutions based on Arm Cortex-M55 and Ethos-U55, and we are very excited and look forward to Arm’s new AI technology, the Ethos-U85 NPU, which will provide the necessary computing performance for our new generation Ensemble series microcontrollers and fusion processors, facilitating future edge AI and visual use cases.”
Steve Tateosian, Senior Vice President of Industrial MCU, IoT, Wireless, and Computing Business at Infineon, stated:“As edge AI use cases become increasingly complex, seizing opportunities in the AI era requires secure, high-performance computing systems. We look forward to making these complex systems possible through the support of the Transformer architecture provided by Arm Ethos-U85 and our long-standing partnership with Arm.”
Corstone-320: New IoT Reference Design Platform Accelerating Deployment of Voice, Audio, and Visual Systems
The Arm Corstone-320 IoT reference design platform integrates Arm’s highest performance Cortex-M CPU—the Cortex-M85, Mali™-C55 ISP, and the new Ethos-U85 NPU, providing the necessary performance for a wide range of edge AI applications such as real-time image classification and object recognition, or enabling voice assistants with natural language translation capabilities on smart speakers. This reference design platform encompasses software, tools, and support, including Arm’s virtual hardware. The combination of hardware and software features of this platform will enable developers to start software development work before the physical chip is ready, thereby accelerating product progress and shortening time-to-market for increasingly complex edge AI devices.
Image: The new IoT reference design platform Corstone-320
Combines leading Arm embedded IP, software, tools, and support,
Accelerating the deployment of voice, audio, and visual systems
Disclaimer: Arm, Ethos, Corstone, Cortex and Mali are registered trademarks or trademarks of Arm Limited (or its subsidiaries).
Source: | Arm Technology Academy
Copyright belongs to the original author. If there is any infringement, please contact for deletion.
END
关于安芯教育
安芯教育是聚焦AIoT（人工智能+物联网）的创新教育平台，提供从中小学到高等院校的贯通式AIoT教育解决方案。
安芯教育依托Arm技术，开发了ASC（Arm智能互联）课程及人才培养体系。已广泛应用于高等院校产学研合作及中小学STEM教育，致力于为学校和企业培养适应时代需求的智能互联领域人才。