Product of the Week: Premio’s RCO-6000-CFL AI Edge Inference Computer

April 04, 2022

Sponsored Story

Product of the Week: Premio’s RCO-6000-CFL AI Edge Inference Computer

New and evolving hardware technologies are making real-time AI inferencing at the edge a reality. These include fluid edge-to-cloud connectivity, expansive networking capabilities, and sustained computational performance at efficient power consumption levels, all of which are designed into Premio’s RCO-6000-CFL AI Edge Inference Computer.

The RCO-6000-CFL is a rugged edge AI computing solution for Industrial IoT applications such as factory automation, machine vision, and fog/edge cloud networking. It leverages an innovative dual-module architecture that allows automation, transportation, and other engineers in critical industries make use of a fanless industrial PC on top of a carrier board housed in an extruded aluminum enclosure that supports additional compute, NVMe storage drives in addition to the 2.5” SATA drive bays already present, and M.2 networking and acceleration cards.

Based on 8th or 9th generation Intel® Core™  (Coffee Lake), Pentium®, or Celeron® processors and the Intel® Q370 chipset, the RCO-6000-CFL offers up to eight CPU cores in its standard configuration. Not only that, but an enhanced bottom configuration called the “EdgeBoost Node” includes an NVIDIA GeForce® RTX 2060 Super Graphics engine that leverages the graphics company’s Turing GPU architecture with 2176 CUDA cores, which combines with the Intel® CPU cores to deliver a balanced mix of parallel compute power and serial determinism.

Up to 64 GB of 2x 260-pin 2400/2666 MHz DDR4 SO-DIMM memory to keep machine learning inferencing data streaming between sensors, deep learning models, and the RCO-6000-CFL’s extensive computing performance. Hot-swappable NVMe “canister bricks” can also be integrated to streamline data management and redundancy.

Premio’s RCO-6000-CFL AI Edge Inference Computer in Action

One often-overlooked aspect of an edge AI inferencing platform is connectivity, as more often than not AI models are trained elsewhere and the data and decisions made locally must be shared with other enterprise systems. This information feedback loop means reliable communications with the outside world will be required at certain points during a machine’s operating lifecycle.

The RCO-6000-CFL’s LGA1151 socket hosts the Intel Q370 chipset that manages high-speed I/O and networking technologies to ensure low-latency, responsive communications at the edge. The RCO-6000-CFL interfaces the chipset manages include:

These connectivity options include, but are not limited to:

  • 5x USB 3.1 Gen 1
  • 4x USB 3.1 Gen 2
  • 1x PCIe x16 and 1x PCIe x4
  • 1/10 GbE PoE LAN via 2x RJ45 and 2x M12 Connectors
  • WiFi 6 ( via integrated Intel® Wireless-AC MAC

There’s also a USB 3.2 header capable of 5 Gbps transfer speeds supported independent of the chipset.

Additional wireless connectivity options include external socket support for multi-SIM Cellular 4G/ LTE and dual-SIM 5G module, and expansion via 2x full-size Mini PCIe slots, while embedded connectivity is present in the form of 6x RS-232/422/485 (2x internal), 16x isolated DIO interfaces, and support for up to six display interfaces over 1x HDMI, 1x DVI-I, 1x DVI-D, and 3 DisplayPort ports.

The multimedia support continues with 1x micin and 1x line-out for audio, which is backed by a Realtek ALC888S codec. All in all, there are plenty of pathways to get data into the system’s powerful edge AI computational and storage resources.

There’s also a TPM 2.0 onboard for enhanced platform security, and a software programmable watchdog ensure safe operation even in harsh industrial edge environments. Over current, over voltage, and reverse polarity protections are obviously built in.

Getting Started with the RCO-6000-CFL AI Edge Inference Computer

All of those features are packed into a combined 261 mm (D) x 240 mm (W) x 168 mm (H) package that weighs just 11 kg Despite its rugged aluminum housing, the RCO-6000-CFL can be mounted on walls and comes pre-certified to FCC Class A. It’s also capable of running in -25ºC to 60ºC temperatures with up to 65W of processing horsepower.

Less surprising is the platforms’ EMC conformance to EN 50155 and 50121-3-2 or 20G shock and 3 Grms vibration rating.

If you’re interested in getting started with the RCO-6000-CFL, you’re in luck, because it supports both Windows 10 and Linux 5.x operating environments. The modular edge inferencing computer can also be used with open-source tools like the Intel® OpenVINO toolkit that optimize deep learning algorithms for CPU and GPU targets, making your transition to machine intelligence that much more seamless.

The RCO-6000-CFL is available now, and more information can be found at Other resources are available below.