Steatite edge AI systems offer real-time visual inference results without the need for high-bandwidth cloud connectivity.
Based on Nvidia CUDA accelerators including Jetson, GeForce, Tesla and Quadro, along with Intel Movidius VPUs and Arria FPGAs.
Real-Time Inference – Running AI models on local hardware results in near-instant inference results.
Power Efficient –Specialist AI acceleration hardware and optimisation techniques helps reduce power consumption.
Frameworks – Regardless of the AI framework used, we can help to get your model running in a live environment.
Harsh Environment – Edge AI inference is ideal for use when shock, vibration, temperature, water or dust can be a problem.
Connectivity – Inference at the edge reduces bandwidth requirements, but we can offer 4G/5G, WiFi & LAN connectivity.
Take a look at some of our AI systems and accelerators below, or contact one of the team to learn more.
Power efficient AI-at-the-edge inference systems based on FPGA, Jetson or Movidius accelerators.
Low-latency FPGA & Movidius based Inference acceleration cards for AI edge computing.
• MXM GPU Module Expansion
• Up to 7 Independent Displays
• 10 Gigabit Ethernet Option
• 8th/9th Gen Intel Core CPU
• Compact & Fanless Design
• Flexible I/O Expansion
• -40°C to 70°C Operating Temp
The GM-1000 is a rugged AI / GPU computing platform supporting embedded MXM GPU expansion. It has all the features required for a compact, reliable, and high-performance computing system for . . .
• Nvidia / AMD GPU Support
• Intel 8th/9th Gen Core CPU
• Up to 7 Display Outputs (Q370)
• Brushed Aluminium Front
• Up to 32GB DDR4 Memory
• Available For At Least 5 Years
The GW-1000 is a compact, high performance graphics workstation offering support for a range of Nvidia and AMD graphics cards. Ideally suited to applications requiring multi-head graphics output, or for . . .
• M.2 AE Key Form Factor
• 1 x Intel Myriad X VPU
• 4.5W Power Consumption
• OpenVINO Toolkit Support
• -20°C to 60°C Operating Temp
• AI Edge Computing Ready
The Mustang-M2AE-MX1M.2 AE-key card includes one Intel® Movidius™ Myriad™ X VPU, providing an flexible AI inference solution for compact size and embedded systems.
• M.2 BM Key Form Factor
• 2 x Intel Myriad X VPUs
• 7.5W Power Consumption
• OpenVINO Toolkit Support
• -20°C to 60°C Operating Temp
• AI Edge Computing Ready
The Mustang-M2BM-MX2 card included two Intel® Movidius™ Myriad™ X VPU, providing an flexible AI inference solution for compact size and embedded systems.
• Mini PCIe Form Factor
• 2 x Intel Myriad X VPUs
• 7.5W Power Consumption
• OpenVINO Toolkit Support
• -20°C to 60°C Operating Temp
• AI Edge Computing Ready
The Mustang-MPCIE-MX2 card included two Intel® Movidius™ Myriad™ X VPU, providing an flexible AI inference solution for compact size and embedded systems.
• PCIe x 2 Form Factor
• 4 x Intel Myriad X VPUs
• 15W Power Consumption
• OpenVINO Toolkit Support
• Multiple Cards Supported
• -20°C to 60°C Operating Temp
The Mustang-V100-MX4 is a PCIe Gen 2 x 2 card included 4 Intel® Movidius™ Myriad™ X VPU, providing an flexible AI inference solution for compact size and embedded systems.
• 2 x Intel Myriad X VPUs
• Low Power Design
• OpenVINO Toolkit Support
• M.2 2280 B+M Key Form Factor
• Complies With CE/FCC Class A
• AI Edge Computing Ready
The EGPA-I201 enhances all vision inference applications such as facial recognition, vehicle registration plate recognition, and many other machine vision applications. Thanks to its low power consumption and high performance, . . .
• MXM 3.1 Type B Form Factor
• 2048 NVIDIA CUDA Cores
• 6.4 TFLOPS SP Performance
• 16GB GDDR5 Memory
• 192GB/s Memory Bandwidth
• 100W Maximum Power
Meeting the needs of embedded, ruggedised, and mobile system builders, the EGX-MXM-P5000 utilises Quadro Pascal architecture to deliver superior graphics and computing performance.
• MXM 3.1 Type B Form Factor
• 1280 NVIDIA CUDA Cores
• 3.9 TFLOPS SP Performance
• 4GB GDDR5 Memory
• 168GB/s Memory Bandwidth
• 75W Maximum Power
Meeting the needs of embedded, ruggedized, and mobile system builders, ADLINK’s EGX-MXM-P3000 Embedded MXM GPU Module is specifically purposed to accommodate form factors incompatible with conventional PCI Express cards, and . . .
• MXM 3.1 Type A Form Factor
• 768 NVIDIA CUDA Cores
• 2.3 TFLOPS SP Performance
• 4GB GDDR5 Memory
• 96GB/s Memory Bandwidth
• Supports Up to 4 FHD Displays
ADLINK’s EGX-MXM-P2000 Embedded MXM GPU Modules features advanced NVIDIA Quadro GPU with NVIDIA Pascal™ Architecture technology in MXM 3.1 Type A form factor. The EGX-MXM-P2000 has 768 NVIDIA CUDA cores . . .
• MXM 3.1 Type A Form Factor
• 512 NVIDIA CUDA Cores
• 1.8 TFLOPS SP Performance
• 4GB GDDR5 Memory
• 96GB/s Memory Bandwidth
• Supports Up to 4 FHD Displays
The EGX-MXM-P1000 features advanced NVIDIA Quadro GPU with NVIDIA Pascal™ Architecture technology in MXM 3.1 Type A form factor. The EGX-MXM-P1000 has 512 NVIDIA CUDA cores and a peak single-precision . . .
Copyright Steatite Ltd / 2019 / All Rights Reserved
