AIO-1688JD4

16T AI Mainboard

AIO-1688JD4 is powered by the SOPHON octa-core AI processor BM1688, delivering up to 16 TOPS of INT8 computing power. It supports
16-channel decoding+10-channel encoding+16-channel AI analysis and video input from six sensors. This mainboard provides mainstream
programming frameworks and the SOPHON SDK, a one-stop deep learning development toolkit. Its complete, easy-to-use toolchain features
easy deployment of algorithm applications and a low cost of algorithm migration. The device can be applied to AI servers, edge AI boxes,
industrial PCs, smart IP cameras, AIOT, and intelligent security, improving various industries through AI

The private deployment of large language models

16 TOPS AI processor BM1688

This mainboard is equipped with the BM1688 SOPHON AI processor, featuring an octa-core ARM Cortex-A53 with a max
frequency of 1.6 GHz. With an integrated TPU, a neural network acceleration engine, it delivers 32T@INT4 peak computing power,
16T@INT8 peak computing power, 4T@FP16/BF16, and 0.5T@FP32. With support for mainstream programming frameworks,
this device can be widely used in AI inference and computer vision

Powerful multi-channel video AI performance

It supports up to 16-channel H.265/H.264 1080P video decoding, 10-channel H.265/H.264 1080P video encoding, and 16-channel
1080P HD video processing (decoding + AI analysis). This makes it ideal for various AI application scenarios such as facial
detection on video streaming, license plate recognition, smart cities, and more

Professional security ISP with
powerful image processing

This professional security ISP is capable of processing video input from six sensors and features hardware-based machine vision algorithms.
It supports HDR, 3D noise reduction, 3A, image edge enhancement, dehazing, and dynamic contrast enhancement—these functions
significantly improve image quality. Additionally, it provides computer vision algorithms such as
fisheye lens unwarping, image stitching, and binocular fusion

Strong network communication capability

This mainboard features two Gigabit Ethernet, and offers expansion capabilities for WiFi/BT modules via the M.2 E-KEY (2230)
interface, supporting WiFi 6 and BT 5.0. It also provides 4G LTE expansion (Mini PCIe) and 5G (M.2),
enabling higher network communication speeds

Extensive connectivity

A wide range of interface options includes MIPI-CSI, MIPI-DSI, HDMI 2.0, SATA 3.0, USB 3.0, USB 2.0, Type-C (Debug), RS485, RS232, CAN, SPI,
I2C, Line in, Line out, M.2 (5G), Mini PCIe (4G), M.2 (WiFi), and M.2 (SSD). These interfaces facilitate the connection of peripheral devices,
enabling different product applications across various fields

Abundant resources

We offer SDKs, tutorials, technical documentation, and development tools to streamline
and improve the development process

A wide range of applications

The mainboard is widely used in AI servers, edge AI boxes, industrial PCs, smart IP cameras,
AIOT, intelligent security, and more

AI servers
Edge AI boxes
AIOT
Industrial PCs
IP Cameras
Security

Specifications

Specifications
Basic Specifications SOC

SOPHON BM1688

Octa-core ARM Cortex-A53 @ 1.6GHz

TPU

The built-in SOPHGO neural network acceleration engine TPU, 32T@INT4 peak computing power,

16T@INT8 peak computing power, 4T@FP16/BF16 computing power, and 0.5T@FP32 computing power

Mainstream neural network architectures: Caffe, Pytorch, TensorFlow(Lite), ONNX and MXNet

ISP

Support time-division multiplexing processing for video input from up to 6 sensors,

with maximum performance support: dual 8M@30 SDR, 8M@30 HDR + 4M@30 HDR, or 8K @15 SDR.

VPU

Video decoding: H.264 / H.265 decoding (max performance: 1920 * 1080@480FPS or 3840 * 2160 @120FPS)

Video encoding: H.264 / H.265 encoding (max performance: 1920 * 1080@300FPS or 3840 * 2160 @75 FPS)

Image codec: JPEG/MJPEG Baseline encoding and decoding (max JPEG codec resolution: 1080P@480 FPS)

RAM

8GB LPDDR4(4GB/8GB/16GB optional)

Storage

32GB eMMC (32GB/64GB/128GB/256GB optional)

Storage
Expansion

1 * TF card, 1 * M.2 (M.2 SATA PCIe/NVME SSD 2242/2260/2280), 1 * SATA 3.0 (SATA 3.0 SSD expansion)

Power

DC 12V (5.5mm * 2.1mm, supporting 9V~24V wide input voltage)

OS

Linux OS (Ubuntu and Debian)

Power
Consumption

Normal: 8.4W (12V/700mA), Max: 30W (12V/2500mA)

Dimension

122.89mm * 85.04mm * 22.7mm

Environment

Operating temperature: -20℃~60℃

Storage temperature: -20℃~70℃

Storage humidity: 10%~90%RH (non-condensing)

Interfaces Network

Ethernet: 2 * RJ45 (1000Mbps)

WiFi: WiFi/BT module expansion available through M.2 E-KEY (2230), supporting WiFi 6 and BT 5.0

4G: 4G LTE expansion available through Mini PCIe

5G: 5G expansion available through M.2, multiplexed with USB3.0 (1)

Video Input

3 * MIPI CSI DPHY(1*4 lanes or 2*2 lanes,30Pin-0.5mm)

Video output

1 * HDMI2.0(4K@60fps)

1 * MIPI DSI DPHY(1*4 lanes or 2*2 lanes,30Pin-0.5mm)

Audio Input

1 * 3.5mm audio jack, supporting MIC recording, CTIA standard

SATA

1 * SATA 3.0 (SATA 3.0 SSD expansion)

USB

2 * USB3.0 (Max: 1A)

Other

1 * Type-C (USB 2.0/DEBUG), 1 * FAN (4Pin-1.25mm), 1 * SIM card

1 * double-row pin header (2*10-20PIN-2.0mm): USB 2.0, SPI, 2*I2C, Line in, Line out, GPIO

1 * Phoenix connector (2*4Pin, 3.5mm pitch): 1 * RS485, 1 * RS232, 1 * CAN 2.0