Aetina unveils MXM module powered by Hailo-8 AI inference processor
Aetina, an edge device manufacturer and AI solution provider, has developed an MXM module featuring a Hailo-8 inference processor. According to the company, the MXM 3.1 module supports high-mobility applications with high performance, lower latency and high throughput demands. Some examples include autonomous driving vehicles/machines and virtual fence systems.
The Aetina AI-MXM-H84A module can deliver up to 26 tera operations per second and a total AI performance of 104 tera operations by leveraging four Hailo-8 processors. The device comes with a comprehensive software development kit with AI development tools. It also supports fine-tuning customization services for AI inference model integration.
“By collaborating with Hailo, we are excited to deliver more edge computing solutions, helping our partners create or adopt AI in different verticals and industries by leveraging the ASIC hardware,” said Jackal Chen, the senior product manager at Aetina.
Since it measures only 82.0 x 105.0 mm, it can be easily integrated into a wide range of embedded systems to handle heavy AI workloads, the company says. It can provide high-speed deployment of neural networks and deep learning models on edge devices. Further, the AI module is efficient, with a typical power consumption of 25W.
The Hailo-8 inference edge AI processor can be used as a standalone processor as well as a co-processor. The easy hardware integration with no additional external memory makes it easy for embedded manufacturers to adopt this inference processor chip. Hailo-8 adopts the company’s novel structure-driven dataflow architecture to enable the processing of the FHD stream in real-time without downsampling.
“Aetina’s MXM modules represent another step forward in making advanced edge AI applications more accessible and easier to use,” stated Gary Huang, the general manager of Greater China business at Hailo. “We are grateful for the close cooperation between Aetina and Hailo and look forward to creating more innovative products supporting high-performance AI applications.”
In September, in response to the rising demand for computer vision applications, Aetina released four new Jetson Orin Nano system-on-module-based edge AI computers. As part of the DeviceEdge family, Aetina’s edge AI computers provide both embedded and standalone computer capabilities.
Zededa provides edge infrastructure to real estate software company Switch Automation
Marvell, Nokia upgrade ReefShark SoC chipset for 5G
Article Topics
Aetina | autonomous vehicles | chip | edge AI | Hailo
Comments