Neousys Technology, an industry-leading provider of rugged embedded systems, announces a new ruggedized edge AI inference platform, Nuvo-7166GC, powered by NVIDIA® T4 GPUs for high-performance AI inference at the edge. The ability to process localized data guarantees response time and can dramatically reduce the data volume uploaded to the cloud.
The NVIDIA edge computing platform spans from the power-efficient NVIDIA Jetson Nano to full-rack NVIDIA T4 servers. Nuvo-7166GC is a leading ruggedized AI inference platform, comprised of Neousys patented Cassette module technology providing optimized cooling for NVIDIA T4 GPUs to ensure stable system operation in harsh environments. In the Cassette module, there is also an additional PCIe x16 slot for a second add-on card installation for application flexibility.
Taking advantage of NVIDIA T4 GPUs, Neousys Technology’s Nuvo-7166GC is designed for advanced inference acceleration applications such as voice, video, image and recommendation services. It offers up to 65 TFLOPS of FP16 compute and 130 TOPs in INT8 for real-time inference applications.
The system is encased in a ruggedized industrial-grade enclosure measuring just 240mm (W) x 225 mm (D) x 111 mm (H). Unrestricted from rack mount or tower size dimensions, Nuvo-7166GC has versatility and deployment advantages over traditional server designs where one can:
Set it up for machine vision and object inspection
The system can combine GbE/ PoE, USB 3.1 Gen2/ Gen1 interface and optional CameraLink/ CoaXPress frame grabber for image acquisition and deep learning-based vision applications.
Embedded it into a medical diagnostic instrument
The compact dimensions make it ideal for medical applications to harness the NVIDIA T4 GPU’s performance for deep learning-based medical imaging to provide diagnostic insights. Featuring abundant I/O technologies to boost overall system flexibility, functionality and performance, the NVIDIA T4 powered Nuvo-7166GC differentiates itself from other server designs. It is an ideal edge AI inference platform for advanced applications.