MicroSys Electronics announced today its partnership with AI (Artificial Intelligence) chipmaker Hailo to launch its miriac AIP-LX2160A embedded platform hosting up to 5 integrated Hailo-8 AI accelerator modules. The new edge server-grade AI solution enables high-performance and scalable AI inference capabilities at the edge.
The new, application-ready AI platform offers industries a high bandwidth and power-efficient solution at the edge, benefiting a wide range of applications in Industry 4.0, automotive, heavy machinery, and more, the company said.
Powered by the NXP QorIQ Layerscape LX2160A high-throughput processor technology, the miriac AIP-LX2160A can integrate multiple advanced Hailo-8 AI accelerators and offers best-in-class processing performance and deep learning capabilities of up to 130 tera-operations per second (TOPS).
The combined solution delivers exceptional AI computing performance across multiple standard NN benchmarks, including over 6000 Frames Per Second (FPS) on Resnet-50, over 5000 FPS on Mobilenet-V1 SSD and close to 1000 FPS on YOLOv5m.
The automotive-grade Hailo-8 is said to outperform other available AI processors for edge computing with up to 26 tera-operations per second (TOPS) at a typical power consumption of 2.5 W.
The embedded platform, combining the Hailo-8 AI processor with MicroSys’s Arm Cortex NXP Layerscape platforms, give customers the benefit of highly efficient AI implemented into their connected edge appliances for situational awareness and predictive maintenance analytics, the company said in a release.
Typical markets for this new application-ready bundle include low-power IIoT and Industry 4.0 edge servers for predictive maintenance, collaborative robotics, video surveillance servers in infrastructures with distributed cameras, communication servers for autonomous vehicles in logistics and agriculture, and heavy equipment for construction, as well as edge servers in trains where multiple GigE Vision camera streams are analysed with AI for increased safety and security.
