Industrial GPU Computer

Powering Edge AI Analytics
with GPU-Accelerated Computer Vision

Premio’s portfolio of x86 Industrial GPU Computers provides a diverse selection of GPU accelerated solutions to cater to a spectrum of edge AI workloads. As Edge AI applications continually become more prevalent and advanced in Industry 4.0, so will the demand for rugged edge solutions to reliably process and consolidate complex AI workflows in real-time.

  • Tailored solutions for various ranges of Edge AI workloads
  • Supports NVIDIA Professional GPUs
  • UL Certified for Safety and Reliability
  • Tested & Validated GPU List
  • Super-Rugged & Semi-Rugged computing solutions
 

Dedicated GPU Support

Real-time Processing

Low Latency

Optimized Data Bandwidth

Scalable & Cost Effective

Industrial GPU Computers for Edge AI Workloads

Our portfolio of industrial GPU computers accommodates various edge AI workloads. From standard to high-performance computing, Premio has an edge computing solution that can best suit the deployment application.

RCO-6000 Series: AI Edge Inference Computer

Low-Profile GPU Support

This industrial GPU computer takes a modular approach by utilizing EDGEBoost technologies. In specific EDGEBoost Node configurations, it offers support for a low-profile GPU to enable various edge AI workloads, along with optional hot-swappable data storage in NVME or SATA.

Key Features:

  • Low-profile GPU
  • PCIe Gen 4
  • Modular EDGEBoost Technologies

Key Applications:

  • People & Vehicle Counting
  • Traffic Flow Analysis
  • Middle-Mile Delivery

VCO-6000 Series: Machine Vision Computer

Dual-GPU Support (Full-Height, Full-Length)

Streamline complex AI workloads with support for dual-GPU configurations. This high-performance industrial GPU computer supports full-height, full-length (FHFL) GPUs for unrestricted edge AI processing.

Key Features:

  • Dual-GPU (FHFL)
  • PCIe Gen 4
  • Hot-swappable NVMe Storage Bays

Key Applications:

  • Quality & Defect Detection
  • Predictive Maintenance
  • Digital Twin & Sensor Fusion

KCO-3000 Series: 3U Fanned Industrial Computer

Dual-GPU Support

This industrial GPU computer offers a COTS approach for rapid time-to-market while delivering maximized edge AI performance with dual-GPU support. Unlike the VCO, the KCO-3000-RPL is only compatible with mid-sized GPUs up to 8.5” (215.9mm) or smaller.

Key Features:

  • Dual-GPU (up to 8.5")
  • PCIe Gen 5
  • 3U Rack Mountable

Key Applications:

  • Industrial Automation Robotics
  • Anomaly Detection & Monitoring
  • Predictive Maintenance

KCO-2000 Series: 2U Fanned Industrial Computer

Low-Profile GPU Support

Purpose-built to seamlessly integrate into existing or new OEM/ODM systems for speedier time-to-market. Leveraging its small form factor, this industrial GPU computer can manage edge AI applications with support for a low-profile GPU.

Key Features:

  • Low-profile GPU
  • PCIe Gen 5
  • 2U Short Depth Chassis

Key Applications:

  • Industrial Automation Robotics
  • Quality & Defect Detection
  • Predictive Maintenance

Featured Industrial GPU Computers:

2U Industrial Fanned Computer 3U Industrial Fanned Computer Machine Vision Computer AI Edge Inference Computer
KCO-2000-RPL
KCO-3000-RPL
VCO-6000-RPL
RCO-6000-RPL
Use Case: Semi-Rugged Semi-Rugged Super-Rugged Super-Rugged
System Cooling Active Cooling Active Cooling Active Cooling Passive Cooling (Fanless)
Performance Acceleration
Expansion Options - PCIe Gen 5 Expansion
- Low-profile GPU
- 1x Hot-swappable SATA SSD
- PCIe Gen 5 Expansion
- Dual-GPU
- PCIe Gen 4 Expansion
- Dual-GPU (FHFL)
- Up to 4x Hot-Swappable SATA/NVMe SSDs
- PCIe Gen 4 Expansion
- Low-profile GPU
- EDGEBoost Nodes
- EDGEBoost I/O
- Up to 8x Hot-Swappable SATA/NVMe SSDs
Durability & Ruggedness
Operating Temp. 0 °C to 60 °C 0 °C to 60 °C -25°C to 70°C -25°C to 70°C
Shock & Vibration 1 Grms
With SSD: 25G
1 Grms
With SSD: 25G
With HDD: 1 Grms
With SSD: 3 Grms / 50G

IEC60068-2-64:2008
Designed to comply with MIL-STD-810G Method 514.7 Procedure I
With HDD: 1 Grms
With SSD: 5 Grms / 20G

IEC60068-2-64:2008
Designed to comply with MIL-STD-810G Method 514.7 Procedure I
Certifications UL (Pending), CE, FCC UL (Pending), CE, FCC UL 62368 Ed. 3, CE, FCC Class A UL 62368 Ed. 3, CE, FCC Class A
See KCO-2000-RPL See KCO-3000-RPL See VCO-6000-RPL See RCO-6000-RPL

Download Our Solution Guide For Embedded Computer and Industrial Touch Display Solutions

Comparing GPU Performance

Support VCO-6000-RPL, KCO-3000-RPL

Support RCO-6000-RPL, KCO-2000-RPL

Model Name RAM CUDA Cores TDP Port Interface Active Cooling Slots Compatibility
NVIDIA T1000 8G 896 50 4x mDP PCIe 3.0 x16 Yes 1
NVIDIA RTX A2000 12G 3328 70 4x mDP PCIe 4.0 x16 Yes 2
NVIDIA RTX 4000 SFF 20G 6144 70 4x mDP PCIe 4.0 x16 Yes 2
NVIDIA RTX A4000 16G 6144 140 4x mDP PCIe 4.0 x16 Yes 1
NVIDIA RTX 4070 12G 5888 200 1x HDMI, 3x DP PCIe 4.0 x16 Yes 2

How Industrial GPU Computers Streamline Edge AI Workloads

Industrial GPU Computers can accelerate edge AI and machine learning workloads with the help of the parallel processing power, CUDA, and tensor cores of GPUs. These features allow for multiple tasks to be operated simultaneously and with mixed precision, which means that calculations are dynamically adapted to accelerate throughput while preserving accuracy.

Key Technologies:

  • Parallel Processing Architecture
  • Tensor Cores
  • CUDA (Compute Unified Device Architecture)

Our Partners

Premio is part of NVIDIA Partner Network that provides in-depth knowledge into the latest technologies powering machine learning and artificial intelligence. Our rugged edge computing solutions are validated to support enterprise performance acceleration cards (GPUs) from NVIDIA's portfolio of compute, virtualization, and visualization products.
Learn More ...

Premio is part of PNY's Pro Partner network of distribution that provides access to NVIDIA's Professional GPUs. PNY is a global technology leader within the OEM, consumer and channel electronics markets. Established in 1985, PNY celebrates its 30th year of business excellence serving consumers, system integrators, OEMs, and B2Bs.
Learn More ...

FAQ

An Industrial GPU computer is a rugged, edge computing solution that utilizes a dedicated GPU (Graphics Processing Unit) card for computational intensive applications requiring advanced image processing and machine vision/intelligence. In addition to a performance boost from the GPU, an industrial GPU computer will also have all the key requirements of what makes up an industrial computer, such as support for high shock/vibration ratings, wide operating temperatures, and over voltage/reverse polarity protection. The industrial GPU computer is transforming how factory automation and smart manufacturing verticals are utilizing computational analysis and machine intelligence in real-time.

  • Computer / Machine Vision
  • Production line inspection
  • Medical imaging
  • Metrology imaging
  • Facial recognition
  • Navigation
  • Autonomous vehicle
  • Pathfinding
  • Automation
  • Factory automation
  • Pathfinding
  • Artificial Intelligence
  • Deep learning
  • Machine learning

Both GPUs (Graphics Processing Unit) and CPUs (Central Processing Unit) have many traits in common; they are both specifically designed microprocessors to handle various tasks. The key difference is HOW they process these tasks. In a computer, the CPU is often referred to as the brain of the system; it is the central processing unit which handles all computing tasks. A GPU, while similar to a CPU, is engineered specifically to process or render graphics. As such, a GPU can work in conjunction with the CPU to help it offload graphics intensive tasks, while freeing up the CPU for other non-graphics related jobs.

Nvidia GEFORCE GTX

The CPU is ideal for serialized, generic tasks, which makes it well-suited for common business or productive applications such as Word, Excel, or a web browser. The number of cores in a given CPU is limited, up to 28 cores for the latest Xeon server class CPU. Conversely, the GPU comes with hundreds or even thousands of cores, which is designed for parallelized, specific tasks; the GPU is optimized for intensive computational applications such as image processing or AI.

As an example, let's examine the fundamentals of image processing. The 4K image of the clock consists of 8.2 million discrete pixels (4K resolution is 3,840 x 2,160, which gives us 8.2 million pixels). From a high level, in order to process the image, we will need to perform some type of computation to each individual pixel. If this task is given to a CPU, with its limited number of cores, the processing time will take very long, as the CPU does not have enough cores to handle the task in parallel. The GPU, which its thousands of specialized cores, can complete the task up to 50-100X faster due to its parallel architecture. This makes the GPU the optimal microprocessor to handle tasks that require parallelism with a high degree of computation.

We begin with a proven embedded system that is engineered to withstand extreme shock and vibration, along with a wide operating temperature range; the inclusion of an industrial class GPU will enable the system to operate reliability in industrial/manufacturing sectors, with the GPU handling AI or image processing applications that require massive parallelism.

One way to compare the performance of various GPUs is by their TFLOP rating. TFLOP stands for "teraflop," which is a measurement of the GPU to performance one trillion floating point operations per second. By adding a GPU to an industrial system, another critical variable we need to consider is the TDP (Thermal Design Power) rating of the GPU. The TDP rating tells us the maximum heat, in watts, generated by the GPU when operating at maximum capacity. This is key in determining the operational temperature range of an industrial GPU system, which should be optimally in the range of -25C to 60C for factory/manufacturing averse conditions. By looking at the Performance versus TDP chart below, we can see the direct correlation between these two values: as the GPU performance increases, so does its corresponding TDP value. When designing an industrial GPU system, there is a fine balance to strike between GPU performance versus overall system operating temperature.

The difference between a GPU and CPU are their computing architecture and task specializations. CPUs are responsible for sequential processing, meaning that it utilizes a smaller number of cores for managing tasks and intensive single-thread applications. GPUs, however, have significantly more cores and leverage parallel processing for simultaneously processing multiple tasks. CPUs are responsible for running the operating system and managing applications while GPUs process machine learning workloads and rendering graphics.

Learn more about CPU vs GPU vs TPU

The key differences between Nvidia’s professional and consumer GPUs are: workload optimization, memory configuration, and power-to-performance balancing. Professional Nvidia GPUs, such as Quadro, are catered towards processing AI workloads with relatively lower TDP, while consumer Nvidia GPUs, like RTX, are focused on general-purpose processing and can have much higher TDP.

Learn more about edge-focused GPUs in our comprehensive GPU guide for industrial computers

Industrial GPU computers work with edge AI by providing the necessary components to enable and process complex workload applications.

First, these computers are purpose-built to withstand the rigorous conditions and environments at the edge. Factory floors and outdoor environments are generally not suitable for consumer desktops as they will eventually fail. Premio’s industrial GPU computers are UL certified, meaning they have undergone thorough testing and validation to ensure safety standard compliance, and select computers have compliance with MIL-STD-810G for shock and vibration.

Secondly, Industrial GPU computers utilize a heterogeneous computing approach to leverage optimized performance through a mixture of specific components, like the CPU and GPU. GPUs are efficient and effective at parallel processing, while the CPU’s strengths are in sequential processing.

Third is IoT connectivity. There are multiple IoT devices in the form of cameras, sensors, and such that need to connect to a centralized computer to process all the incoming data. Industrial GPU computers provide compatible I/O to connect and even power IoT devices.

By balancing both rugged reliability, processing power, and IoT connectivity, industrial GPU computers help drive complex edge AI workloads at the rugged edge.

Learn more about how industrial GPU computers enable Edge AI here

Yes, industrial GPU computers are machine vision systems that deliver the needed processing power and I/O connectivity to enable machine vision applications.