NSR7936 Kunpeng Dual AI Server

Industrial ICT & Industry Terminals Industrial Board / Host

Contact manufacturer

NSR7936 Kunpeng Dual AI Server

Overview

A dual-socket rack server for AI inference, model fine-tuning, on-premise knowledge bases, HPC and simulation rendering. It leaves room for full-size dual-width GPU cards, paired with Kunpeng processors and onboard memory for concentrated compute deployments.

When selecting, focus on GPU count and power budget, PCIe expansion, drive bays and network scaling. The unit targets model deployment, training / fine-tuning and HPC nodes. The site can supply the unit and compute solution; configuration is confirmed with the manufacturer.

ApplicationsAI inference, model fine-tuning, local knowledge bases, HPC, simulation and rendering.

Specifications

  • Rack server with two Kunpeng 920-Smart modules
  • supports four dual-width full-size GPU cards (350W+), up to four Ascend accelerators
  • six PCIe Gen4 slots
  • 8/12 3.5" drive bays.

Comments

No comments yet. Be the first to comment.