SOPHON Intelligent Computing Server SG6-06-A22

SOPHON SG6-06-A22 intelligent computing server can be equipped with 18 Sophon third-generation TPU chip BM1684, INT8 performance up to 316.8TOPS, support 17280fps 1080P high-definition video hardware decoding and 900fps 1080P high-definition video hardware encoding.

Strong computing power

Support up to 316.8TOPS of INT8 performance
or 39.6TFLOPS of FP32 performance

Ultra high performance

576 channels of full HD video decoding
36 channels of full HD video hardware encoding
Video and picture decoding resolution up to 8K or more

Fast porting

Support Caffe/TensorFlow/PyTorch/
MXNet/Paddle Lite and other mainstream deep learning frameworks

Flexible expansion

Storage supports 12 3.5”/2.5” front hot-swappable hard drives, Slots and 2 2.5” rear hot-swappable hard drives, 8 PCIE3.0 expansion slots, 24 DDR4 Memory slot, the maximum capacity supports 3TB

High compatibility

Intel dual-socket scalable Xeon processor platform
Supports RESTful/IPMI management interface

Wide range of applications and rich scenes

Adapt to multiple intelligent algorithms of many partners, and perfectly support intelligent analysis applications such as face recognition, vehicle recognition, item recognition, video structuring, behavior analysis, etc.

Easy to use and convenient, full stack efficient

BMNNSDK (BITMAIN Neural Network SDK) one-stop toolkit provides a series of software tools such as the underlying driver environment, compiler, inference deployment tool and so on. Easy to use and convenient, covering the model optimization, efficient runtime support and other capabilities required for the neural network inference stage, providing easy-to-use and efficient full-stack solutions for deep learning application development and deployment. BMNNSDK minimizes the development cycle and cost of algorithms and software. Users can quickly deploy deep learning algorithms on various AI hardware products of Fortune Group to facilitate intelligent applications.

Support mainstream programming framework


Performance parameter

AI performance






Video encoding capabilities

H.264: 1080P @17280fps / H.265: 1080P @17280fps


Video encoding resolution

8192 * 8192 / 8K / 4K / 1080P / 720P / D1 / CIF

Video encoding capabilities

H.264: 1080P @900fps / H.265: 1080P @900fps

Video encoding resolution

4K / 1080P / 720P / D1 / CIF

Video transcoding capability

Maximum 324 channels (1080P to CIF)

Picture decoding capability (JPEG)

14400 sheets/sec @1080P

Picture decoding maximum resolution

32768 * 32768

Video memory

216GB LPDDR4x 4000MHz


ARM A53 144 core @2.3GHz


2 Intel Xeon Scalable 2nd CPU


24 * DDR4 DIMM 2933MHz, up to 3TB


12 2.5/3.5" SATA/SAS hard disks


2 x 1G LAN port / 4 x USB3.0 / 1 x VGA / 1 x COM / 1 x MGMT

Power supply

Two 1200W platinum power supplies support 1+1 redundancy mode

Heat dissipation

4 fans

Operating temperature

10°C - 35°C

Operating system


Dimensions and weight

2U (438mm*88mm*730mm), net weight 30.2kg