• atlas-800-inference-3000-banner-pc

    Atlas 800 Inference Server (Model: 3000)

    Powerful computing, ultra-high efficiency


  • Overview
  • Features
  • Specifications
  • Resources

Atlas 800 Inference Server (Model: 3000)

Powered by the Ascend processor, the Atlas 800 inference server (model: 3000) supports up to 8 Atlas 300I inference cards to provide powerful real-time inference. It is widely used for AI inference in data centers.

Rapid Deployment

Superior computing

• Supports 8 Atlas 300I inference cards to meet the inference requirements in multiple scenarios; 640-channel real-time HD video analytics (1080p 25 FPS)
• Runs on the 64-core Kunpeng 920 processors to unlock powerful computing for application acceleration

Superior perf./watt

• Provides an AI computing platform with high efficiency, low power for inference scenarios, fully leveraging the multi-core, low-consumption advantages of Kunpeng
• Atlas 300I runs at only 67 W, fueling the AI server with faster computing and higher performance per watt


Parameters Atlas 800 Inference Server
Model: 3000
Form Factor 2U AI server
Processor 2 Kunpeng 920 processors
Processor Memory 32 DDR4 DIMM slots, up to 3200 MT/s
AI Accelerator Card Up to 8 Atlas 300I inference cards
Local Storage 25 x 2.5'' SAS/SATA drives
12 x 3.5'' SAS/SATA drives
8 x 2.5'' SAS/SATA + 12 x 2.5'' NVMe
RAID RAID 0, 1, 10, 5, 50, 6, or 60
PCIe Up to 9 PCIe 4.0 PCIe ports, among which one is a PCIe slot dedicated for screw-in RAID controller card, and the other 8 are for plug-in PCIe RAID controller cards
Power Supply 2 hot-swappable 900 W or 2000 W AC PSUs, supporting 1+1 redundancy
Fan Module 4 hot-swappable fan modules, supporting N+1 redundancy
Operating Temperature 5°C to 40 °C (41°F to 104°F)
Dimensions (H x W x D) 86.1 mm x 447 mm x 790 mm (3.39 in. x 17.60 in. x 31.10 in.)