- June 28, 2020
June 29, 2020 - Dynics, a U.S. based manufacturer of industrial-grade computer hardware, visualization software, network security, network monitoring and software defined networking solutions, announced the XiT4 Inference Server.Artificial Intelligence (AI) is increasingly being integrated into modern manufacturing to improve and automate processes, including 3D vision applications. The XiT4 Inference Server, powered by the NVIDIA T4 Tensor Core GPU, is a fan-less hardware platform for AI, machine learning and 3D vision applications.
The XiT4 is powered by an NVIDIA T4 Tensor Core GPU which features 320 NVIDIA Turing Tensor Cores and 2560 NVIDIA CUDA cores for critical processes that can benefit from processing large sets of data. The T4 supports a variety of AI frameworks for training and inference and is supported by several SDKs, such as , for data analysis and visualization. The XIT4 also features an Intel Core i7-8700T CPU, 32GB DDR4 RAM and up to terabytes of SSD storage capacity. Operating system choices are Linux Ubuntu 20.04 LTS and Windows 10 Enterprise 2019 LTSC. Additionally, other Linux distributions are available upon request.
The overall dimensions including mounting tabs are 19.861”H x 6.25” W x 13.553” D; mounting tabs are included for panel mount or on the machine. Standard I/O configuration includes three USB 3.0, four 1Gb Ethernet ports and DVI video; it also includes an illuminated power button. Additional USB ports available as well as options for 10 Gb Ethernet, 1Gb PoE+, and fiber optic. The system uses an L-coded cable for 24vdc input supply power.
The XiT4 has an IP50 rating for indoor use and a passively cooled, fan less design; an IP65 option is also available. This allows for 24/7 runtime on the plant floor at operating temperatures of 0-50 deg. C. This on-premise solution allows for data collection and computing at the edge complementing current or future cloud solution.