Advancing Industrial AI with AMD Embedded+ Architecture

Single PCB with Embedded Processors & Adaptive SoCs

AMD’s Embedded+ single-board computer architecture combines the company’s Ryzen embedded processors and its Versal AI SoC onto a single PCB, giving system designers and ODMs powerful computing solutions that offer reduced latency between hardware and software. With the ability to distribute workloads across x86 and ARM CPUs as well as across AI engines and FPGA fabrics, AMD’s single-board solution is ideally suited for deployment in edge AI systems that target visual inspection and defect detection in demanding machine vision applications.

Derived from AMD’s Ryzen CPU, the AMD Ryzen embedded processor provides reduced power consumption while maintaining scalable x86 processor instructions designed to manage heavy workloads and “always on” applications.

The Versal AI SoC brings to market a powerful dual-core ARM Cortex-A72 processor ideal for running Linux applications and for delivering low-latency AI inference coupled with the ability to accelerate the application from sensor to processing to real-time control.

Sensor Friendly

The wide-ranging compatibility that Embedded+ offers with various sensor types (RGB/mono, time of flight, lidar, ultrasonic, GPS, and others), as well as its broad interface support for MIPI, SubLVDS/LVDS, and SLVS-EC, promotes broader design flexibility and ease of integration with industrial and embedded PCs.

Distribute Processing

Utilizing multiple processing technologies, AMD Embedded+ provides integrators with the ability to delegate processing tasks to the processor best suited for each task. Integrators can leverage low-latency FPGA fabric for communication and processing sensor data while directing inference tasks to purpose-designed AI engines, ensuring that the AMD Ryzen x86 CPU is available for handling operating system and general application processing.

Accelerate Time to Market

Pre-integration combines the AMD Embedded+ solution with a Mini-ITX form factor motherboard, enabling ODMs to quickly and cost-effectively develop computing solutions that can be easily integrated and deployed. Available expansion capability extends support for I/O boards that boost connectivity to include additional memory, PCIe 3.0 x4 M.2, and a SATA port for connecting hard drives and solid-state drives.

Enhancing Performance in Machine Vision and Edge AI

The single-board architecture of AMD’s Embedded+ allows ODMs to develop Mini-ITX form factor motherboards (170 x 170 mm), so machine vision integrators can capitalize on a diverse set of processing tools using industrial PCs and other computing platforms that are ideal for deploying at the edge.

With the proliferation of multi-sensor vision and inspection applications growing steadily, AMD Embedded+ provides the resources necessary for low-latency fusion among different sensors and sensor types, facilitating powerful data processing that can then be fed to AI processing engines in real time. The result is a zero-distance image acquisition and processing solution that can alleviate the work of central processing in applications where bringing the capability to the edge is preferred.

Using the available expansion slots, system integrators and device manufacturers can integrate support for frame grabber–based interface protocols as well as expand support for Ethernet-based vision protocols beyond gigabit data rates. This flexibility enables scalable solutions for the most demanding automated inspection and defect detection applications. Thanks to the x86 architecture of the AMD Ryzen Embedded+ CPU, main system processing tasks can be performed in Windows or Linux environments, separate from the image, data, or inference processing tasks handled using ARM processing or available FPGA fabric. The result is a highly adaptable and configurable compute solution.

AMD Technologies Powering Embedded+

AMD Versal AI Edge Series

AMD’s Versal AI Edge Series accelerates the whole application from sensor to AI to real-time control. With support for multiple sensor-level interfaces such as GSML and Ethernet-based vision protocols, the system is built for integration into complex sensor-driven systems. Further, Versal AI is adaptive, allowing developers to easily advance their sensor fusion and AI algorithms within the portfolio of scalable, performance-oriented devices, all while supporting industrial functional safety and cybersecurity standards.

AMD Ryzen Embedded R2000 Series

Target Applications for Embedded+

Robotics

Healthcare

Machine Vision

Networking & Control

Smart Systems

Now Accepting Orders!

Embedded+