T-Chip Group
Intelligent hardware website
Innovate intelligent hardware and solutions.
Server Business Website
Server products and efficient solutions.
Consumer Business Website
StationPC NAS products and services.

CSB1-N4AGXOrin AI Server

1100 TOPS Computing Power

· Private deployment of AI models  

· Multiple deep learning frameworks  

· 8K video decoding, 4K video encoding  

· Supports 3.5-inch SATA3.0 HDD/SSD  

· 4-port 10 Gigabit SFP+ network interface

· Standard 1U rack server size

· Includes aBMC management system

· Highly integrated server design

Manual Wiki Download

Llama

Qwen

Stable Diffusion

CSB1-N4AGXOrin
CSB1-N4AGXOrin AI Server

CSB1-N4AGXOrin can be configured with up to 4 NVIDIA Jetson AGX Orin (64 GB) compute modules, delivering a maximum computing power of 1100 TOPS (INT8). It supports the private deployment of mainstream AI models and various deep learning frameworks. Equipped with four 10-gigabit network ports, it also allows for expansion with SATA 3.0 hard drives.

Private Deployment

Supports private deployment of mainstream large language models (such as Gemma, Llama, Qwen, etc.), vision large models (such as EfficientVIT, SAM, TAM, etc.), image generation models (such as Flux, Stable Diffusion, etc.), and other large-scale models.

Deep Learning Frameworks

Supports the Ollama local large model and ComfyUI graphical deployment framework; compatible with multiple deep learning frameworks accelerated by cuDNN, including PaddlePaddle, PyTorch, TensorFlow, MATLAB, MXNet, and Keras; supports custom operator development and Docker containerization management technology.

8K Decoding ＆ 4K Encoding

Supports decoding of up to 1 channel of 8K@30fps, 3 channels of 4K@60fps, and 11 channels of 1080P@60fps video, as well as encoding of 2 channels of 4K@60fps and 8 channels of 1080P@60fps video, meeting the demands of various AI application scenarios.

3.5-inch SATA 3.0 HDD/SSD

Features a 3.5-inch (or 2.5-inch) drive bay supporting SATA 3.0 HDD/SSD expansion, enabling seamless capacity scaling to terabyte-level storage. With hot-swappable capability for rapid drive replacement, it fulfills integrated deployment requirements for file management, data backup, video surveillance, and similar applications.

4-port 10 Gigabit SFP+

Equipped with four 10-gigabit SFP+ network ports for high-speed, stable communication to meet high-bandwidth application scenarios. A dedicated BMC management network interface is included to separate management and data traffic, ensuring network security and reliability.

Standard 1U Server Size

Features a standard 1U rack-mounted chassis design with flexible deployment capabilities, fully compatible with most data center cabinet types. The robust construction incorporates 6 high-efficiency cooling fans to ensure 7×24 continuous stable operation.

aBMC Management System

Equipped with an advanced aBMC intelligent management system that enables real-time monitoring, software/hardware configuration, troubleshooting, anomaly alerts, system upgrades, and remote maintenance. Supports secondary development for customized requirements.

Efficient and Low-Cost

Server highly integrates computing units, storage, universal serial bus, network controller, power controller, and sensors into one, providing SDK all-in-one deep learning development toolkit, underlying driver environment, compiler, inference deployment tool, and a series of software tools to reduce users' procurement, development, and operation costs.
1100 TOPS Computing Power

Configured with up to four NVIDIA Jetson AGX Orin (64GB) compute nodes, delivering a peak performance of 1100 TOPS (INT8) to provide robust computing power for AI and deep learning applications.

Interfaces
Specifications

CSB1-N4AGXOrin

Technical Specifications

Server form

1U rack-mounted computing power server

Architecture

ARM architecture

Number of nodes

4 distributed computing nodes + 1 control node

Compute nodes

12-core 64-bit processor NVIDIA Jetson AGX Orin (64GB), main frequency up to 2.2GHz

Video encoding

2×4K@60fps, 4×4K@30fps, 8×1080P@60fps, 16×1080P@30fps

Video decoding

1×8K@30fps, 3×4K@60fps, 7×4K@30fps, 11×1080P@60fps, 22×1080P@30fps

Control nodes

Octa-core 64-bit processor RK3588, main frequency up to 2.4GHz, the highest computing power is 6TOPS

AI computing power

1100TOPS (275T × 4, INT8)

RAM

64GB LPDDR5 × 4

Storage

64GB eMMC × 4

Storage expansion

3.5-inch/2.5-inch SATA3.0 SSD hard drive slot × 1 (Supports hot swapping; BMC can directly operate the hard drive, and computing child nodes can indirectly access the hard drive through the network sharing method provided by BMC)

Power

550W AC power supply (Input: 90V AC~264V AC, 47 Hz~63 Hz, 8A; Hot swappable not supported)

Power consumption

Normal: 135W, Max: 400W

Fan module

6 high-speed cooling fans

Physical Specifications

Size

494.0mm(L) × 469.3mm(W) × 44.4mm(H)

Installation requirements

IEC 297 Universal Cabinet Installation: 19 inches wide and 800 mm deep and above Retractable slideway installation: The distance between the front and rear holes of the cabinet is 543.5mm~848.5mm

Full weight

Server net weight: 8.8kg, total weight with packaging: 11kg

Environment

Operating Temperature: 0ºC ~ 40ºC, Storage Temperature: -40ºC ~ 60ºC, Operating Humidity: 5% ~ 80%RH (non-condensing)

Software Specifications

BMC

The BMC management system is integrated with the web-based management interface, supporting Redfish, VNC, NTP, monitoring advanced and virtual media, and the BMC management system can be redeveloped.

Large language models

Support the privatization deployment of ultra-large-scale parametric models under the Transformer architecture, such as Deepseek-R1 series, Gemma series, Llama series, ChatGLM series, Qwen series, Phi series and other large language models.

Visual large model

Support the privatization deployment of large visual models such as EfficientVIT, NanoOWL, NanoSAM, SAM and TAM.

AI Painting

Supports the private deployment of Flux, Stable Diffusion, and Stable Diffusion XL image generation models.

Deep learning

Support Ollama local large model deployment framework and ComfyUI graphical deployment framework. Support for a variety of deep learning frameworks supported by cuDNN acceleration, including PaddlePaddle, PyTorch, TensorFlow, MATLAB, MXNet, Keras. Support custom operator development and Docker containerization management technology.

Interface Specifications

Internet

4 × 10G Ethernet (SFP+), 1 × Gigabit Ethernet (RJ45, MGMT is used as BMC management network)

Console

1 × Console (RJ45, BMC debug serial port, baud rate 115200)

Display

1 × HDMI (maximum resolution 1080P, BMC management display)

USB

2 × USB3.0 (The lower USB is USB3.0 OTG, and the BMC can be upgraded OTG by using a USB flash drive)

Button

1 × Reset, 1 × UID, 1 × Power

Other interfaces

1 × RS232 (DB9, baud rate 115200),1 × RS485 (DB9, baud rate 115200)
Customization

Firefly team, with over 20 years of experience in product design, research and development, and production, provides you with services such as hardware, software, complete machine customization, and OEM server.

Hardware Solution

Software Solution

Turnkey Solution