en.Wedoany.com Reported - Forlinx Embedded has launched the FET3572-C core board based on the Rockchip RK3572 processor, integrating a 4 TOPS NPU and an octa-core heterogeneous architecture, delivering AI and multimedia processing capabilities for scenarios such as intelligent security, commercial displays, and edge computing.

As AIoT deepens its integration, the demand for AI computing power and multimedia processing in edge devices continues to rise. Forlinx Embedded, a strategic partner of Rockchip, has introduced the FET3572-C core board featuring the RK3572 processor, equipped with a 4 TOPS (INT8) hardware-accelerated NPU that supports INT4/INT8/INT16 mixed-precision computing and is compatible with mainstream frameworks such as TensorFlow and PyTorch, enabling direct deployment of edge AI inference tasks. The core board is based on a 2×Cortex-A73 plus 6×Cortex-A53 octa-core heterogeneous architecture, manufactured using an 8nm process, delivering over 100% performance improvement compared to the previous mid-range platform while reducing typical power consumption by more than 50%. It scores over 310,000 on Antutu v10, with standby power consumption below 10mW.

In terms of multimedia processing, the core board supports 8K decoding and 4K encoding, is compatible with mainstream and open-source formats, integrates a 12 MP ISP for precise color reproduction, and supports 5 camera inputs and dual independent displays (4K@60fps plus 2K@60fps). For connectivity, it offers PCIe 2.1, dual Gigabit Ethernet, CAN-FD, and I2C industrial bus interfaces, with native support for LPDDR5/5X memory. The software ecosystem supports Linux and Android systems, and comes with driver source code, development manuals, and technical examples. The core board is suitable for application scenarios such as intelligent security, industrial control, edge computing, and in-vehicle systems.

The FET3572-C core board is now open for pre-orders. For more information, please visit: https://www.forlinx.net/

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com