ASUS Small Form Factor SBCs Suit Industrial Edge AI Applications
May 11, 2023
Some new single-board computers (SBCs) from ASUS IoT up the ante in terms of performance for industrial Edge AI applications.
First, the Tinker Board 3 is designed with a 64-bit Arm quad-core Cortex-A55 processor and includes support for both Linux (Yocto and Debian) and Android OSs. Interfaces include PoE, LVDS, COM and CAN bus, plus M.2 E and M.2 B slots to accommodate wireless or 4G/5G expansion modules. As evidence that “embedded” is the target for the board, ASUS is providing at least seven years of availability. Aimed at industrial applications, the 100- by 100-mm board can operate in temperatures ranging from -40℃ to +85℃.
A relative of the Tinker 3 is the Tinker V, which is based on a RISC-V architecture and fits the compact Pico-ITX form factor. The company claims that this is the first SBC based on the relatively new RISC-V open-source architecture. At 64 bits, it provides the performance needed for higher end industrial IoT (IIoT) applications. Linux Debian and Yocto support is included.
Specifically, the Tinker V is equipped with a Renesas RZ/Five MPU, which incorporates the RISC-V AndesCore AX45MP single-core CPU, running at 1.0 GHz. It is also engineered with an array of peripheral connectors aimed at industrial applications, with GPIO, micro-USB, dual gigabit Ethernet, a pair of CAN bus interfaces, and two RS232 COM ports. It also includes 1 Gbyte of built-in RAM and an optional 16-Gbyte eMMC. Operating temperatures range from -20°C to +60°C.
Finally, the PE6000G system is also aimed at Edge AI applications. It supports up to a 450-W graphics card for high-throughput, demanding GPU-computing applications. Powered by a 12th Gen Intel Core processor and paired with the Intel R680E chipset, the PE6000G is ready for the latest PCI Express (PCIe) 5.0 standard, meeting requirements for real-time AI inferencing and deep learning training at the Edge. It supports up to 64 Gbytes of ECC/non-ECC DDR5 4800 SDRAM, four 2.5-in. SSDs, and one each of M.2 M key (NVMe), M.2 B key (5G NR), and M.2 E key (WiFi 6). It also provides all the connectivity that you’d expect for Edge AI applications.