(858) 613-1818Quote list 0

2U dual socket Intel Xeon 6 compute server

Dual Intel Xeon 6 server in 2U with up to 256 performance cores, DDR5 or MRDIMM memory, 24 drive bays and room for two accelerators or 400 Gb/s adapters.

Illustration of a 2U rackmount server
CPU
2 x Intel Xeon 6, performance cores up to 128 per socket or efficient cores up to 288 per socket
Memory
16 or 32 x DDR5 RDIMM or MRDIMM, up to 4 TB
Storage
2 x M.2 boot, up to 24 x 2.5 inch NVMe or SAS bays
Networking
1 x 400 Gb/s ConnectX-7, 2 x 25 GbE onboard
Expansion
Up to 6 x PCIe Gen5 slots, space for 2 double width GPUs
Power
Redundant PSUs, about 1 kW per node planning figure

Overview

The 2U Intel Xeon 6 server is the flexible CPU node for environments standardized on Intel. Two Xeon 6 processors offer up to 256 performance cores with AMX for CPU side inference, or up to 576 efficient cores for throughput services, with eight DDR5 channels per socket and MRDIMM support on performance core models for the highest memory bandwidth.

The 2U height adds 24 drive bays and six PCIe Gen5 slots, so the same node serves as a data staging server, a database host, a virtualization node or a light GPU server with two double width cards. In GPU clusters we deploy it in the CPU tier at six threads and 24 GB per GPU, and as the head node for job scheduling and user login.

Configuration includes mirrored boot drives, a 400 Gb/s ConnectX-7 adapter, dual 25 GbE and redundant power supplies that accept 100 to 240 V. Burn in covers memory, drives and network ports under sustained load, and the acceptance report with serial numbers ships with each unit.

Use cases

  • CPU tier for GPU clusters
  • Head and login nodes
  • Databases and analytics
  • Virtualization and containers
  • CPU inference with AMX

Specifications

Compute
CPUs2 x Intel Xeon 6, P core models up to 128 cores per socket, E core models up to 288 cores per socket
System memory16 or 32 x DDR5 RDIMM, 8 channels per socket, up to 4 TB; MRDIMM supported on P core models
AcceleratorsBuilt in AMX for AI inference on CPU, QAT and DSA engines
Sizing rule6 CPU threads and 24 GB memory per cluster GPU served
Storage
Boot2 x M.2 NVMe, mirrored
DataUp to 24 x 2.5 inch NVMe or SAS, hot swap
RAIDHardware RAID or HBA
Networking
Fabric1 x 400 Gb/s ConnectX-7 or 2 x 200 Gb/s
Frontend2 x 25 GbE onboard, OCP 3.0 slot for 100 GbE
ManagementDedicated 1 GbE BMC port, TPM 2.0
Physical
Form factor2U rackmount, air cooled
Power suppliesRedundant 1,600 W class PSUs, 100 to 240 V input
Planning powerAbout 1 kW per node without GPUs

Figures describe the system class. Exact configuration, weights and power draw are confirmed on the quotation.

What's included

  • Rails, power cords and cable management arm
  • Burn in and acceptance test report
  • Operating system imaging and firmware baseline on request
  • 3 year warranty, options to 5 years with advance replacement

Compliance

This product is not a data center GPU system and does not require NVIDIA end user registration. Standard export documentation applies. Read about export compliance.

Compare it with the next class up or down

The nearest classes we quote are the 1U dual socket EPYC 9005 server and the 2U four node multi node server. If neither is quite right, tell us the workload and we specify the system between them.

Talk to an engineerRequest a quotation