NVIDIA Studio Certified · NVIDIA GeForce Authorized System Integrator · Intel Gold Partner · AMD Arena Member
X-SERIES · SCALE-UP GPU

GL-X6U-H200

NVIDIA HGX H200

6U scale-up GPU server on NVIDIA HGX H200. One chassis, 2 host platforms: AMD EPYC 9005 and 5th / 4th Gen Intel Xeon. Validated with GenesysBench™ and deployed under GLIDS; AI workloads run on the GLabs AI Stack.

Platform
HGX H200
GPUs
8
Memory
24 DIMM slots
NVMe bays
8×
GL-X6U-H200, scale-up gpu server, overview

Specifications

Pick the host platform. The chassis, power and cooling follow it.

GL-X6U-H200: AMD EPYC 9005 specifications
Form factor6U Rackmount
Processor1+1 Socket SP5 (LGA6096); AMD EPYC™ 9005/9004 (with AMD 3D V-Cache™ Technology) and 97x4 series processors
CPU TDPUp to 500W
Memory12+12 DDR5 DIMM Slots (1DPC); RDIMM up to 128GB each, max. 6400MT/sRDIMM-3DS up to 256GB each, max. 6400MT/s
GPUSupports eight NVIDIA H200 GPUs
PCIe fabricPCIe Gen5 switch
Expansion8 HHHL PCIe5.0 x162 FHHL PCIe5.0 x16
Storage8 hot-swap 2.5" NVMe (PCIe5.0 x4) drive bays from PCIe Switch1 M-key (PCIe3.0 x2 or SATA 6Gb/s), 22110/2280 form factor1 M-key (PCIe3.0 x4 or SATA 6Gb/s), 22110/2280 form factor
Networking2 RJ45 (1GbE), share with front panel1 dedicated IPMI
Power4+4, 80-PLUS Titanium, 3000W CRPS
Cooling21 PWM 80x80 mm fans
Dimensions (D × W × H)930 x 448 x 264.7 mm (36.6" x 17.6" x 10.4")
ManagementDedicated BMC (IPMI)

NVIDIA-Qualified GPUs

  • NVIDIA HGX H200 8-GPU · on 5th Gen Intel Xeon Scalable, AMD EPYC 9005

Exact CPU, memory, drive and network parts are fixed on your quote. GPU and platform names are trademarks of their respective owners.

Configure GL-X6U-H200

Choose the host, memory, storage and networking. Your request goes straight to a Solutions Architect, who replies with a validated build and quote.

  1. 01System and host
  2. 02Processor
  3. 03Memory
  4. 04GPU
  5. 05Storage
  6. 06Networking
  7. 07Software stack
  8. 08Send request
Start configuring

Tell us what you're building.

A solutions architect responds directly — no ticket queue.

Talk to a Solutions Architect