Combined with comprehensive I/O, flexible expansion, and Advantech's integrated expertise, including Edge AI SDK and WISE-Edge Developer Architecture (WEDA), these platforms accelerate the deployment of AI in industrial automation, medical imaging, transportation, and interactive outdoor urban systems, enabling faster time to market, lower development risk, and long-term reliability.
The AMD Ryzen™ AI Embedded P100 Series processors feature up to 12 cores and 24 threads, based on the latest “Zen 5” core architecture. They support the AVX-512 instruction set, including VNNI and BFLOAT16, providing high compute density and broad software compatibility for the most demanding edge applications. Integrated AMD RDNA™ graphics, utilizing the AMD ROCm™ software ecosystem, can support up to four 4K displays at 120Hz with a hardware codec accelerator. This enables advanced visualization for machine vision applications and interactive human-machine interfaces (HMIs). Combined with the new AMD XDNA™ 2 NPU architecture, the platform delivers up to 80 total TOPS for AI inference acceleration. This heterogeneous architecture streamlines the management of complex AI workloads to deliver optimal performance, low latency, and energy efficiency in both Windows and Linux development environments. Supporting up to 128GB of DDR5-5600 RAM, PCIe® Gen 4 x8, NVMe, HDMI 2.1, DisplayPort 2.1, and USB 4, these processors provide excellent scalability and flexibility. TDP options range from 15W to 54W, with variants supporting temperatures from -40°C to 105°C, ensuring energy efficiency without compromising performance.
Advantech offers a comprehensive portfolio of Ryzen™ AI Embedded P100 series processors for diverse EIoT applications. The MIO-5380 3.5" single-board computer (SBC) combines a compact and robust design with an expansion module (MIOe series) that supports external GPUs via MXM, additional LAN ports with PoE, and PCIe expansion via the MCIO connector. EdgeBMC enables out-of-band remote management even if the OS is unresponsive. Building on this flexibility, the SOM-6874 COM Express Compact Type 6 module supports up to 18 PCIe lanes and multiple I/O ports, including 2.5 GbE and USB 3.2, with a QFCS thermal design that prevents throttling down to 60°C and motherboard reference designs to accelerate application development. For larger-scale integration, the AIMB-2211 Mini-ITX motherboard adds robust connectivity with four independent 4K displays, seven USB ports plus an optional 100W USB Type-C port, a PCIe 4.0 x8 slot, and a M.2 NVMe and a wide-range 12–24V continuous power input simplify deployment in compact systems.
Motherboards with AMD Ryzen™ AI Embedded P100 series processors are validated with Windows 11 LTSC and Ubuntu LTSC and are complemented by a suite of value-added software to accelerate development and deployment. To expedite AI application development, the Advantech EdgeAI SDK provides a cross-platform toolkit that leverages the capabilities of the AMD Ryzen™ AI processor, enabling efficient deployment of pre-trained models and accelerating time to market. Additionally, Advantech DeviceOn serves as an intelligent device management platform, offering remote monitoring, predictive maintenance, and centralized control to simplify large-scale edge AI operations. At the system level, the Advantech SUSI API provides standardized interfaces for hardware control and monitoring. This facilitates access to GPIO, watchdog timers, and thermal and power management to improve stability and performance. Efficiency. Complementing these software tools, Advantech's Custom Design services offer expert advice and technical support, from motherboard reference designs to thermal optimization and compliance testing, to help customers reduce development complexity, accelerate deployment, and ensure long-term reliability.
