6U 8-GPU AI Inference Server S6685

Processor Dual 4th/5th Gen Intel® Xeon® Scalable processors
Up to 64 cores per CPU
GPU Support Up to 8 × PCIe Gen5 full-height, full-length, dual-width GPUs
NVIDIA RTX 5090 / RTX 6000-class GPU support
Memory 32 × DDR5 RDIMM, up to 5600 MT/s
Maximum capacity: 8 TB
Storage Module1: 8 × 3.5" SATA/SAS + 4 × 2.5" NVMe drives
Module2: 12 × 3.5" SATA drives

    Scenarios

  • LLM Inference (Llama, DeepSeek)

  • Image Gen, Video Codec, Super-Res

  • Scientific Computing & HPC

  • Performance

    Up to 350W TDP CPU support

    PCIe Gen5 architecture for high-throughput GPU workloads

  • Reliability & Security

    Multi-zone thermal design with intelligent fan control

    PFR (Platform Firmware Resilience) support

    TPM security module support

  • Flexible Configuration

    Modular design for flexible drive and I/O configurations

    Broad PCIe expansion support

AI Server Specification —— Intel CPU Base

Specifications 6U rack mounted
CPU 2 × 5th Gen / 4th Gen Intel ® Xeon ® Scalable processors (Eagle Stream)
Memory 32 x DDR5 5600MT/s RDIMM, maximum capacity 8 TB
Storage Module1: 8 × 3.5" SATA/SAS + 4 × 2.5" NVMe drives
Module2: 12 × 3.5" SATA drives
RAID support Optional support for RAID0 , 1 , 1E , 10 , 5 , 50 , 6 , 60
Support Cache super capacitor protection, provide RAID state migration, RAID configuration memory
PCIe Expansion Rear:8 x FHFL dual width PCIe 5.0
Front: 3 x FHHL PCIe 5.0
Switch:6 × FHHL PCIe 5.0
I/O interface 4 × USB
1 × BMC management port
2 × VGA
1 × COM
BMC Support IPMI 2.0, Redfish protocol, provide remote management functions such as KVM, realize hardware monitoring and management, precise control, and support deep customization for customers
Power 4 or 8 platinum power supplies
Support N+1/N+N redundancy
Operating system Mainstream operating systems such as Microsoft Windows Server, Red Hat, Ubuntu, CentOS, etc
Dimensions 482mm × 855mm × 263.7mm

Let’s get started
with us.

For furthur info & support, Contact us.