The COM-HPC specification was finally approved on February 24th and is available for download on the PICMG homepage. Almost simultaneously, Congatec is providing a starter set for the standard. The starter set for COM-HPC is optimized for modular system designs that use the latest high-speed interface technologies such as PCIe Gen4, USB 4.0 and ultra-fast up to 2x25 GbE connectivity as well as the integrated MIPI-CSI-Vision functionality. It is based on Congatec's PICMG COM-HPC Computer-on-Module Conga-HPC/cTLU, which is equipped with 11th generation Intel Core processor technology (codenamed Tiger Lake).
"This new high-end embedded module generation is aimed at system engineers who are developing broadband networked edge devices that are emerging as part of the development of the industrial IoT," says Martin Danzer, Director Product Management at Congatec. Target markets include medical technology, automation, transportation and autonomous mobility, as well as image processing-based inspection and video surveillance systems.

The various Ethernet configuration options of the starter set range from 8x 1GbE switching options and 2x 2.5GbE including TSN support to dual 10GbE connectivity. Congatec also offers AI support for MIPI-CSI-connected cameras from Basler, making IIoT and Industry 4.0-connected embedded systems more applicable. AI and inference acceleration can be achieved on the CPU using Intel DL Boost-based vector neural network instructions (VNNI) and on the GPU using 8-bit integer instructions (Int8). Also attractive in this context is the support of the Intel Open Vino ecosystem for AI, which contains a function library and optimized calls for OpenCV and OpenCL kernels to accelerate deep neural network workloads across platforms and thus achieve faster and more accurate results for AI inferences.

Embedded vision and AI

With the low-power SMARC 2.1 Computer-on-Modules, robust 2 - 6 Watt low-power platforms for vision and AI can be realized.

© Congatec

Congatec also presented new low-power SMARC 2.1 Computer-on-Modules with NXP i.MX 8M Plus processor for industrial edge analytics, embedded vision and artificial intelligence (AI). With its machine and deep learning capabilities, the new ultra-low-power Conga-SMX8-Plus module enables industrial embedded systems to see and analyze their environment for situational awareness, visual inspection, identification, monitoring and tracking as well as gesture-based, touchless machine operation and augmented reality.

The technical highlights of the Arm Cortex-A53-based quad-core processor platform include the integrated Neural Processing Unit (NPU) for AI computing power and the Image Signal Processor (ISP) for parallel real-time processing of high-resolution images and video streams from the two integrated MIPI-CSI camera interfaces. The SMARC module ecosystem - such as application-ready 3.5-inch carrier boards and Basler camera and AI software stack support - complements the product launch for a fast proof-of-concept. Vertical markets for these credit card-sized low-power vision and AI modules are expected to be found everywhere from smart farming and industrial manufacturing to retail and from transportation to smart cities and smart buildings.

Energy-saving 24/7 systems

The energy consumption of the AMD Ryzen Embedded V2000 processors is very low compared to their computing power. The Conga-TCV2 computer-on-modules are therefore predestined for solar-powered stationary, mobile and autonomous systems

© Congatec

The Conga-TCV2 is also being launched on the market based on AMD Ryzen Embedded V2000 processors. With twice the performance of the previously launched AMD Ryzen Embedded V1000, the COM Express Compact Computer-on-Module module sets a new performance-per-watt benchmark, delivering its maximum performance improvement in 15-watt TDP designs. Compared to modules with AMD Ryzen Embedded V1608B processors, Conga-TCV2 modules deliver a performance increase of 97% (V2516) to 140% (V2718) with up to 8 cores. Thanks to the new 7 nm Zen 2 cores, single-core performance has also increased by between 25% and 35%. This makes the new modules a perfect candidate for performance increases in 24/7 connected and fanless embedded systems used on a wide variety of industrial edges. Typical applications include multifunctional industrial edge gateways, digital signage systems, gaming terminals and infotainment platforms. With up to 40% more GPU performance[2] for up to 4x 4k60 graphics at 15 watts and comprehensive GPGPU support, medical multi-display systems for imaging in the operating room as well as machine vision and machine learning systems are further target markets.