RK3588+12 GPU edge computing server design solution, supporting 120 channel high-definition video

RK3588+12 GPU edge computing server design solution, supporting 120 channel high-definition video

The high-density server is a edge computing product with high performance and high computing power. It is equipped with an intelligent visual depth learning processor BM1688, and 12 computing units are optional. Supports intelligent analysis of 192 high-definition videos (hardware decoding) and hardware encoding of 120 high-definition videos; Supports multiple Ethernet ports, HDMI, USB, and storage expansion. It can flexibly respond to high concurrency business scenarios such as public security, transportation, industrial parks, production, and warehousing, and is equipped with diverse algorithms to empower various industries with intelligence.

Type

main parameter

Product Name

High density server

Product model

XM3588-BM1688

AI computing power

192TOPS (INT8) peak computing power

Encoding and decoding capability

Video decoding: H.264&H.265: 16x1080P@30fps , supports 8192 * 8192/8K/4K/1080P/720P/D1/CIF

Video encoding: H.264&H.265: 10x1080P@30fps Supports 8K/4K/1080P/720P/D1/CIF

Number of nodes

12 computing nodes (up to 96 ARM cores)+1 control node

processor

Computing Node: BM1688 Core (A53 * 8) 64 bit processor with a maximum clock speed of 1.6GHz

Control node: RK3588 octa core (4 x Cortex-A76+4 x Cortex-A55) 64 bit processor, with a maximum clock speed of 2.4GHz

Memory

8GB LPDDR4 x 12 (number of compute nodes), 32GB eMMC x 8 (number of compute nodes), 2.5-inch SATA 3.0 hard drive x 2

network

Gigabit Ethernet Gigabit RJ45 network port x 5 (1 control node port, 1 regular network port), SFP+x 2, 4G/LTE/5G network (optional)

display

HDMI 2.0 ， 4K@60Hz (Main processor core board display)

USB

1 x USB 3.0 HOST, 1 x USB 2.0 HOST1 x Type-C (device BMC unit debugging test)

Fan

5 high-speed cooling fans

Power Supply

300W AC power supply (input: 100V AC~240V AC)

Linux BMC

Integrate BMC management system, provide web-based management interface, BMC management system can be further developed

Deep Learning Framework

TensorFlow / PyTorch / Paddle / Caffe / ONNX / MXNet / DarkNet

Specification and size

Standard 1U rack mounted server: 490mm x 390mm x 44.4mm

working temperature

0 º C-50 º C working humidity 10%~90% RH (no condensation)