TGU Series accelerated by new Xe GPU
DFI's new-gen TGU Series embedded computing solutions are powered by 11th generation Intel® Core™ SoC processors designed for Edge AI vision computing. In the all-new Intel® Iris® Xe architecture, the high-performance, low-power SoC integrates with the GPU and CPU, and is combined with AI acceleration to make a big progress in the edge AI computing performance and handle complex computing applications, such as computer vision or high workloads, faster and more accurate.
Multiple Specs for Diverse Demands
The new industrial motherboards and embedded computer modules of TGU series are equipped with 11th Gen Intel® Core™ processors in a wide range of specifications from SBC, to Mini-ITX, to COM Express Compact, and to COM Express Mini. They meet the needs of industrial control, surveillance, traffic analysis, and vision computing at the edge.
All-new GPUs in Xe architecture
The Intel® Iris® Xe graphics integrated in the 11th Gen Intel® Core™ SoC have 96 compute units and 16MB of L3 cache, delivering disruptive GPU graphics performance. Paired with DDR4 memory on TGU series motherboards with increased bandwidth to 3200MHz and capacity to 64GB, it enables faster processing of images and video streams, and through minimal jitter drives real-time, efficient processing of multiple IoT tasks and workloads.
More Accurate AI Inference
The new TGU series has received a significant upgrade on the AI architecture. Besides the new support for Deep Learning Boost (DL Boost) and Vector Neural Network Instruction Set (VNNI), which were previously only available on Xeon processors, it also supports graphic card or Intel® Movidius™ VPUs AI accelerator card expansion for additional deep learning and inferencing capabilities.
Broadband, low-latency network transfer
The independent 2.5G LAN provides a faster transmission experience than a typical 1G LAN, and can be combined with the out-of-band (OOB) management system vPro and AMT to optimize and ensure the real-time and seamless collaboration of edge devices. The TGU series also supports broadband 5G, Wifi 6, and other wireless transmissions, giving more flexibility to device deployment beyond the limits of network cables.
Optimized Transfer and Interfaces
The new TGU series features the highest speed PCIe 4.0 transfer signal at present, offering a transmission speed up to 16 GT/s. It benefits the M.2 interface to enjoy the ultra-speedy transfer experience by connecting to either NVMe SSD or DFI proprietary M.2 expansion modules for USB 3.2 Gen1, SATA, COM, LAN, and PSE. The TGU series also offers a wide range of expansion interfaces, including PCIe x4, PCIe x1, M.2 M, B, E keys, SIM, and SATA, etc., supporting a variety of storage and network interfaces to provide more possibilities for edge computing innovation.
3-in-1 Transfer Design
All types of connectors in the near future will soon be united into a single interface – USB Type C. DFI TGU series motherboards offer the future-leading USB Type C interface, integrating three functions including display, data transfer and power supply via a USB hub at once, such as DisplayPort for 4K 60Hz display, LAN port for networking or surveillance, or NVMe SSD for data storage, and simultaneously powering up the USB hub or SSD. This 3-in-1 transfer design makes it easier to integrate the system with the interface of the latest devices to connect to the future.
4 X 4K HDR with high contrast resolution
The TGU Series offers up to four channels of 4K high dynamic range (HDR) independent displays, providing more details of light and dark, images closer to reality, and even realizing ultra-smooth video streams. The high contrast resolution allows every small detail on x-rays and tomograms to be clearly displayed, delivering a more accurate diagnostic assistance to healthcare professionals.
Enhanced IoT Network Security
The TGU Series protects devices from cold boot attacks with the new Total Memory Encryption (TME) feature, and ensures that devices are fully protected against possible external attacks at booting, data storing, or data processing.
