Premio offers modular path to upgradability for edge AI hardware in industrial applications

Categories Edge Computing News  |  Hardware
Premio offers modular path to upgradability for edge AI hardware in industrial applications

Premio Inc. has announced a significant design upgrade for its current line of RCO-6000-CFL AI Edge Inference Computers.

As distributed edge computing solutions shift into more remote, mobile, and unstable conditions closer to IoT sensors, next generation designs require versatile engineering to support real-time processing and inference calculations in conditions prohibitive to traditional PC designs. Premio’s 30+ years of engineering expertise brings an innovative modular approach to industrial-grade compute and system-level design with its “EDGEBoost Nodes.” These new nodes address current demands for powerful performance acceleration needed to push industrial-grade computers to process and analyze large volumes of data for rugged edge computing. Specific computer vision and edge computing applications can benefit from real-time processing capabilities with the latest hardware acceleration technologies by incorporating Premio’s modular AI Edge Inference Computers.

“The design innovation behind our EDGEBoost nodes showcases our industry-leading engineering experience and capability in x86 computing architecture to address environmental issues such as strong vibration, severe temperatures, and the power instability,” Premio’s Product Marketing Director, Dustin Seetoo explains. “With this design upgrade, end-users and system integrators can now scale and future-proof next generation enhancements by leveraging a modular approach to meet their demands in machine learning and data aggregation for AI.”

The RCO-6000-CFL AI Edge Inference Computer series’ unique 2-piece modular design offers tailored performance flexibility from central processors (CPU), graphic engines (GPUs), m.2 accelerators, and even NVMe storage technology. The top computer node provides full I/O access (USB, COM, LAN, Display Ports, DIO) to an industrial fanless PC, while the bottom node is dedicated for new EDGEBoost node models. For a full lineup of available EDGEBoost Nodes, download the product brief.

Key design upgrades support high-speed, low-latency storage NVMe SSDs for storage and m.2 acceleration cards for additional machine learning performance. Premio is first-to-market hot-swappable NVMe canister bricks in a ruggedized embedded system design. Depending on the EDGEBoost node model, a variation of cannister bricks can support up to x4 2.5 inch / 15mm U.2 NVMe SSDs and up to x8 2.5 inch / 7mm U.2 NVMe SSDs available soon. Storage redundancy and data retention is available through hardware and software RAID option. The 2.5 inch drive trays can also support m.2 acceleration cards through a plug and play carrier board. Both the individual drive trays and the cannister bricks are toolless, hot-swappable, and even lockable for data security.

Each EDGEBoost node also includes a software controlled hot-swappable fan to ensure thermal regulation of high-performance acceleration cards from GPU, NVMe SSDs, and M.2 cards. Everything from fan speed, temperature, and programmable fan logic can be monitored. But the most unique feature the software development kit offers is the programable logic that suspends all I/O transmission and read/write operations from the NVMe storage devices to prevent the loss or corruption of data with a click of a button. A dedicated button and LED light provide status indication for when the NVMe cannisters bricks can be safely ejected and undocked for data offload. This feature is pivotal to ensure mission-critical data retention in enterprise applications that rely on data and even streamlines the overall time it takes to swap each NVMe cannister brick for field replacements. Application developers can use the development kit as a building block for their application specific requirements.

Ultimately, by combining NVMe storage performance, GPU Add-In cards, and m.2 acceleration in modular EDGEBoost nodes in its hardware design, the RCO-6000-CFL AI Edge Inference Computer Series is a computing powerhouse fit for the demands of machine learning and real-time inferencing at the edge.

Article Topics

 |   |   | 

Comments

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Edge Ecosystem Videos

Featured Edge Computing Company

Latest News