MPU has built-in vision AI accelerator
Semiconductor manufacturer Renesas Electronics has announced that it has expanded its AI-capable RZ/V Series of microprocessors (MPUs), with a new device that enables AI processing of image data from multiple cameras, offering a new level of highly accurate image recognition for vision AI applications. Equipped with two 64-bit Arm Cortex-A53 cores, the RZ/V2MA MPU is capable of delivering high computing performance with a maximum operating frequency of 1GHz.
The device features a proprietary low power DRP-AI (Dynamically Reconfigurable Processor) accelerator that can process vision AI at 1 TOPS/W (tera operations per second, per watt) class performance. The RZ/V2MA device offers high-speed interfaces such as Ethernet, USB, and PCI Express that allow image input from multiple external cameras.
In addition to the DRP-AI accelerator, the RZ/V2MA includes an OpenCV accelerator that allows rule-based image processing simultaneously. These features, says the company, bring highly accurate image recognition capabilities for machine vision products such as AI-equipped gateways, video servers, security gates, POS terminals and robotic arms.
The RZ/V2MA offers a full suite of development tools to aid vision AI system design. In addition to the existing DRP-AI Translator, the new device adds DRP-AI TVM, which is based on the open-source deep learning compiler Apache TVM technology. While DRP AI Translator is designed to convert AI models into DRP-AI executables, the DRP-AI TVM compiler lets the DRP-AI accelerator work together with the CPU, allowing DRP-AI to convert and generate more AI models than ever before.
As a first phase, says the company, it supports ONNX and PyTorch AI models and plans to support Tensorflow in the future.
“One of the challenges for embedded systems developers who want to implement machine learning is to keep up with the latest AI models that are constantly evolving,” says Shigeki Kato, Vice President of Renesas’ Enterprise Infrastructure Business Division. “With the new DRP-AI TVM tool, we are offering designers the option to expand AI frameworks and AI models that can be converted to executable formats, allowing them to bring the latest image recognition capabilities to embedded devices using new AI models.”
Key Features of the RZ/V2MA include:
- Two 64-bit Arm Cortex-A53 cores with a maximum operating frequency of 1 GHz
- AI accelerator DRP-AI (1 TOPS/W class), achieving 52 fps (frames per second) when executing TinyYoloV3 programs
- OpenCV accelerator for rule-based image processing
- Ethernet, USB, and PCI Express interfaces for image input from external cameras
- Video codecs (H.265 and H.264)
- DRP-AI TVM tool for converting AI models based on TVM technology. Support ONNX and PyTorch formats initially.
- High-speed memory interfaces include LPDDR4 (3200Mbps), USB 3.1 (up to 5Gbps), and PCI Express (2 lanes)
- Available in 15mm square BGA package
The company has developed the “Vision AI Gateway Solution,” which is an AI-based object detection and recognition platform that uses multiple cameras to collect and efficiently transmit data wirelessly. The high-speed processing solution combines the RZ/V2MA MPU with complimentary Renesas products such as power ICs, VersaClock clock generator, and communication modules for Wi-Fi, Bluetooth, and LTE. This solution not only provides flexible connectivity options but has an optimized power supply system and has been tested to accelerate development of robust AI gateway devices.
The RZ/V2MA and development tools are available now.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
