Today I’ll be reviewing the BeagleY-AI open-source single-board computer (SBC) developed by BeagleBoard.org for artificial intelligence applications. It is powered by a Texas Instruments AM67A quad-core Cortex-A53 processor running at 1.4 GHz along with an ARM Cortex-R5F processor running at 800 MHz for handling general tasks and low-latency I/O operations. The SoC is also equipped with two C7x DSP units and a Matrix Multiply Accelerator (MMA) to enhance AI performance and accelerate deep learning tasks. Each C7x DSP delivers 2 TOPS, offering a total of up to 4 TOPS. Additionally, it includes an Imagination BXS-4-64 graphics accelerator that provides 50 GFlops of performance for multimedia tasks such as video encoding and decoding. For more information, refer to our previous article on CNX Software or visit the manufacturer’s website. BeagleY-AI unboxing The BeagleY-AI board was shipped from India in a glossy-coated, printed corrugated cardboard box. Inside, the board is protected by […]
Firefly introduces Rockchip RK3576 SoM and All-in-One carrier board compatible with NVIDIA Jetson Orin Nano and Orin NX modules
Firefly has released a Rockchip RK3576 SoM and development board called the Core-3576JD4 Core Board with a SO-DIMM edge connector and the AIO-3576JD4 carrier board respectively. The core board or the SoM is built around an octa-core 64-bit processor with a Mali G52 MC3 GPU and a 6 TOPS NPU, so it can handle demanding AI tasks while maintaining low power consumption. The AIO-3576JD4 is a full-fledged carrier board with a wide range of on-board interfaces, like dual Gigabit Ethernet ports, MIPI-CSI, HDMI 2.1, USB 3.0, USB 2.0, USB Type-C, a Phoenix connector for serial, dual-row pin headers (SPI, I2C, Line in, and Line out), an M.2 socket for 5G, a mini PCIe for 4G LTE, an M.2 socket for WiFi 6/BT 5.2, and a third M.2 socket for SATA/PCIe NVMe SSD expansion. RK3576 AI SoM and dev board specification Core-3576JD4 specifications SoC – Rockchip RK3576 CPU – Octa-core CPU […]
Qualcomm Networking Pro A7 Elite is a WiFi 7 SoC with a 40 TOPS NPU for AI-optimized traffic and range
The Qualcomm Networking Pro A7 Elite platform is a new wireless networking platform that integrates WiFi 7 with edge AI. The wireless networking platform can deliver up to 33 Gbps of throughput and features an AI coprocessor that provides up to 40 TOPS for AI workloads. AI features are designed to enhance Wi-Fi and networking user experience and include Smart Traffic Classifier, Range Boost, and Networking AIOps. The platform also includes 10Gbps Fiber (XGS-PON), Ethernet, front-end modules, and filters into an integrated platform. All these features are designed to streamline system development and reduce design complexity by integrating key elements—from broadband to antenna—into an integrated platform. Qualcomm Networking Pro A7 Elite platform specification Qualcomm’s Wireless Networking Platform CPU – Unnamed quad-core processor, built on a 14 nm node, and clocked at 1.8 GHz NPU – Up to 40 TOPS (AI coprocessor featuring Qualcomm Hexagon NPU) System Memory – DDR3L/DDR4 Storage […]
reComputer R1000 Raspberry Pi CM4-powered IoT Gateway runs open-source software for Edge AI applications (Sponsored)
Seeed Studio is a Raspberry Pi-approved reseller and device designer. In May 2024, Seeed released their first edge IoT Raspberry Pi-powered gateway and controller, reComputer R1000. It is Raspberry Pi CM4-powered, with AI capabilities when equipped with Hailo NPU. It comes with all the necessary features that a robust and reliable industrial edge IoT device needs, along with rich and versatile interfaces like three RS485 ports, dual Ethernet, and flexible IoT wireless communication (4G, LoRa, Wi-Fi/BLE, Zigbee) and essential functional modules (GPS, UPS, TPM2.0, PoE, SSD, etc.). With Modbus and BACnet supported, it’s a perfect fit for remote access control, especially BMS, BAS, and iBMS. You’ll find more technical details in our previous article about the reComputer R1025, and on Seeed Studio’s product page where you can also purchase the AIoT gateway. As a powerful edge IoT controller and gateway with AI capabilities, it embraces the Raspberry Pi community, is […]
DigiPort is an HDMI computer dongle powered by a Raspberry Pi CM4 (Crowdfunding)
Shivam Goyal, going under the Geeky Tronics name, has developed the DigiPort HDMI computer dongle powered by a Raspberry Pi CM4 system-on-module and designed to be connected directly to the back of an HDMI or through an HDMI cable. Since it does not support MHL, the DigiPort also needs a USB-C power source. You can add a keyboard and a mouse through its two USB 2.0 ports or via Bluetooth and network connectivity is managed through WiFi 5 making it a portable computer ready to use out of the box. DigiPort specifications: Supported System-on-Module – Raspberry Pi CM4 SoC – Broadcom BCM2711 quad-core Cortex-A72 processor @ 1.5 GHz System Memory – 1GB to 8GB LPDDR4-3200 SDRAM Storage – 0GB (CM4 Lite), or 8GB to 32GB eMMC flash Wireless – Dual-band WiFi 5 and Bluetooth 5.0 Storage – MicroSD card for OS when using a CM4 Lite module Video Output – […]
OpenWrt One WiFi 6 router with Filogic 820 SoC launched for $89
The “OpenWrt One/AP-24.XY” is a Filogic 820-based WiFi 6 router board manufactured by Banana Pi whose software is directly managed by OpenWrt developers with assistance from MediaTek. The router was first announced in January 2024, and developer samples became available sometime in April with some early units auctioned away at the OpenWrt Summit which took place in Cyprus on May 18-19. The good news is that the OpenWrt One is now available to anyone on Aliexpress for $89 including a metal enclosure, a PoE module, three antennas, and a power supply. Here’s a reminder of the OpenWrt One router specifications: SoC – MediaTek MT7981B (Filogic 820) dual-core Cortex-A53 processor @ 1.3 GHz System Memory – 1GB DDR4 Storage 128 MB SPI NAND flash for U-boot and Linux 4 MB SPI NOR flash for write-protected (by default) recovery bootloader (reflashing can be enabled with a jumper) Two types of flash devices […]
UnifyDrive UT2 – A compact NAS powered by Rockchip RK3588C with NVMe SSD support and AI-driven storage management (Crowdfunding)
The UnifyDrive UT2 is a compact network-attached storage (NAS) device powered by the Rockchip RK3588C processor. It supports WiFi 6, and Bluetooth, and has multiple ports for connectivity. Equipped with two M.2 2280 slots for PCIe 3.0 NVMe SSDs, it offers up to 16 TB of storage. The device can connect to networks via 2.5 GbE or WiFi and also function as a WiFi access point, allowing file transfers even without an internet connection. A built-in battery provides uninterruptible power supply (UPS) functionality, enabling continued operation when unplugged. Designed for both home and mobile use, the UnifyDrive UT2 offers AI-driven storage management in a portable form factor. Its versatile networking options allow remote access, backups, and on-the-go file transfers. Despite its size, it delivers robust functionality, serving as a full-featured NAS solution in a variety of settings. I didn’t find any significant differences between the Rockchip RK3588 and RK3588C. This […]
Congatec conga-TCR8 AMD Ryzen Embedded 8000 COM Express computer-on-module series delivers up to 39 TOPS of AI performance
Congatec has released a new line of Ryzen Embedded 8000-powered COM Express Type 6 Compact computer-on-modules with the conga-TCR8 module series offering up to 39 tera operations per second (TOPS) for AI inference and 128GB of DDR5-5600 memory with error correction code (ECC) for data-sensitive and data-critical applications. The Ryzen Embedded 8000 processor series comes with either six or eight ‘Zen 4’ cores, an integrated XDNA neural processing unit (16 TOPS), and the AMD Radeon RDNA 3 graphics unit which can be used as a general-purpose graphics processing unit with up to 12 compute units (adding up to 39 TOPS for the SoC). Graphics output is supported by four displays with up to 8K resolution, and peripheral options include six PCIe Gen 4 slots, four USB 3.2 Gen 2 ports, four USB 2.0 ports, and three DisplayPort interfaces. The conga-TCR8 Type 6 modules are intended for high-volume applications that require […]