ATX motherboard facilitates the integration of AI computing
DFI has unveiled a server-grade ATX motherboard, designed for Intel® Ice Lake platform, powered by the 3rd Generation Intel® Xeon® Scalable processors, and equipped with ultra-high speed computing that can support up to 205-W. ICX610-C621A also comes with built-in Intel® Speed Select Technology (Intel® SST), which provides an excellent load balancing between CPUs and multiple accelerator cards to effectively distribute CPU resource, stabilize computation loads and maximize computing power. As a result, it improves the performance by 1.46 times compared to previous generation.
Featuring powerful performance, the server-grade ATX motherboard offers three PCIe x 16, two PCIe x8 slots and one M.2 Key and enables ultra-performance computing, AI workload and deep learning, specifically for high-end inspection equipment, such as AOI, CT, and MRI applications. The ICX610 also supports ECC RDIMM up to 512-GB 3200-MHZ, enhances high end performance for advanced inspection equipment and improves efficiency. For example, this is especially useful for empowering 3D surgery training to develop mature smart surgical technologies.
Regarding the AI inference performance, with the boost from Deep Learning Boost (DL Boost) and Vector Neural Network Instructions (VNNI), the inference for production line and medical inspection equipment has been improved by 1.56 times compared to the previous generation, producing more accurate and reliable automated inspection result.
Through the upgraded 10 Gbps high-speed network, large high-resolution images can be quickly transmitted, accelerating the communication and cooperation between machines (M2M).The ICX610-C621A can help equipment manufacturers and users to further realize true Industry 4.0 automation.
Typical applications include factory automation, AOI flaw detection, AI imaging, healthcare, medical imaging as well as AI diagnostics and assistance
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
