Join the 155,000+ IMP followers

electronics-journal.com

High-Efficiency Embedded AI Module for Industrial Edge Systems

Congatec expands industrial embedded computing with AMD-based COM Express modules for rugged, scalable edge AI applications.

  www.congatec.com
High-Efficiency Embedded AI Module for Industrial Edge Systems

Congatec has introduced a new series of COM Express 3.1 Type 6 Compact Computer-on-Modules (COMs), designated conga-TCRP1, leveraging AMD Ryzen AI Embedded P100 Series processors. The modules target industrial edge AI use cases across transportation, medical, smart infrastructure, robotics, and automation, addressing performance, power efficiency, and ruggedization requirements for embedded computing.

Advancing Edge AI Compute with Integrated Heterogeneous Processing
The conga-TCRP1 modules integrate AMD Ryzen AI Embedded P100 processors combining up to six Zen5 CPU cores, an RDNA 3.5 GPU, and an XDNA2 Neural Processing Unit (NPU). This heterogeneous architecture delivers up to 59 TOPS of combined AI inference performance — up to 50 TOPS from the NPU with the remainder from CPU/GPU compute — enabling real-time processing of machine learning workloads, including compact large language models, without resorting to discrete AI accelerators.

The P100 processor family itself targets embedded AI and graphics workloads with optimized performance and determinism. According to the underlying AMD architecture specifications, the P100 Series combines high-performance “Zen 5” cores with an RDNA 3.5 GPU capable of driving multiple high-resolution displays, and delivers up to 50 TOPS AI inference in constrained embedded environments.

Ruggedization and Scalability for Diverse Industrial Environments
Designed for industrial edge deployments, the conga-TCRP1 modules support an extended operating temperature range from −40 °C to +85 °C on selected variants, accommodating harsh conditions encountered in transportation systems, outdoor smart-city infrastructure, and field robotics. The configurable thermal design power (TDP) range from 15 W to 54 W enables system designers to balance power consumption, thermal constraints, and performance within size, weight, power, and cost (SWaP-C) targets.

Modules in the series provide up to 96 GB of DDR5-5600 memory with optional Error Correction Code (ECC) for mission-critical data integrity, and extensive I/O including PCIe Gen 4, 2.5 GbE networking, multiple USB ports, and storage options via onboard NVMe or SATA interfaces. These interfaces support integration with industrial Ethernet, fieldbus adapters, and radio modules, broadening applicability in automation and control systems.

Software and Development Ecosystem Support
The conga-TCRP1 series supports a range of operating systems pertinent to edge computing, including Microsoft Windows 11, Windows 11 IoT Enterprise, and several Linux distributions. Application-ready variants can be preconfigured with licensed operating systems and optional hypervisor-on-module capabilities to consolidate real-time control, human-machine interface (HMI), AI inference, and IoT gateway functions on a single module. Software building blocks for industrial IoT connectivity facilitate data exchange, remote maintenance, and optional cloud integration.

Application Areas and Technical Benefits
The conga-TCRP1 modules are positioned for embedded use cases where integrated AI inference, graphics, and real-time control are essential. Examples include rugged handheld diagnostic devices, hygienic medical PCs requiring sealed enclosures and passive cooling, machine vision systems in industrial automation, and smart transportation infrastructure with local data processing. The combination of high AI TOPS, scalable memory and I/O, and extended temperature support enables designs that otherwise might require separate processing and acceleration hardware, simplifying system architecture and reducing integration effort.

Relevance to Digital Supply Chain and Automotive Data Ecosystem
In contexts such as smart factories and connected vehicles, the ability to perform local AI inference and heterogeneous compute at the edge supports the digital supply chain by reducing latency, bandwidth usage, and dependency on centralized cloud resources. For automotive data ecosystems, integrated graphics and AI processing facilitate advanced cockpit experiences and real-time sensor data interpretation while meeting robustness requirements for in-vehicle systems.

www.congatec.com

  Ask For More Information…

LinkedIn
Pinterest

Join the 155,000+ IMP followers