The modules are positioned as a versatile solution for high-performance edge computing, AI inference, and graphics-intensive applications.
Congatec has introduced new COM Express Compact Computer-on-Modules powered by AMD Ryzen Embedded 8000 Series processors. These modules, named conga-TCR8, are built with up to eight ‘Zen 4’ cores, an AMD XDNA NPU, and Radeon RDNA 3 graphics, offering high performance for AI-driven applications with processing power up to 39 tera operations per second (TOPS) for AI inference.
The new modules are designed for applications that require a combination of advanced AI, graphics, and computing power, making them suitable for sectors like medical imaging, test and measurement, AI-supported POS/POI systems, and professional gaming. The modules provide long-term availability, ensuring investment security while also offering scalability with a thermal design power (TDP) range of 15 to 54 watts. These modules allow for easy upgrades of existing systems by replacing older modules, improving lifecycle and return on investment (ROI).
Key specifications include
- Up to 8 Zen 4 cores
- Size: 95 x 95 mm (3,74″ x 3,74″)
- Up to 128 GB RAM (optional with ECC)
- TDP ranges from 15W to 54W
- Onboard NVMe SSD (option)
The modules also feature AMD’s XDNA NPU and Radeon RDNA 3 graphics, which support AI tasks and immersive graphics output on up to four 8K displays. For connectivity, the modules offer six PCIe Gen 4 lanes, multiple USB ports, DisplayPort interfaces, and SATA options, with an onboard NVMe SSD available as an option. The company’s ecosystem supports these modules with a range of services, including pre-installed operating systems such as Ubuntu or RT Linux and system consolidation features. This includes a ready-to-use option for custom applications through the aReady.COMs platform.
Additionally, the company offers board support packages, evaluation and carrier boards, cooling solutions, and extensive design-in services to simplify integration and accelerate time-to-market. For more information, click here.