HPE Cray XD670 (8x NVIDIA H100 SXM, Dual 4th Gen Intel Xeon)
Proven 8x H100 training node for production deep-learning pipelines.
We help you choose, configure, and deliver the right system — no obligation.
Configuration at a Glance
Tailored per engagement. Full technical overview below.
Configuration Options
Core specifications for this system. Every component is configurable to your workload — request a quote for a tailored build.
8x NVIDIA H100 SXM, 80GB HBM3 each (640GB total)
Dual 4th Gen Intel Xeon Scalable (Sapphire Rapids)
Up to 32x DDR5-5600 DIMMs
Up to 8x SFF NVMe U.2/U.3 + 2x M.2 RAID
Overview
This HPE Cray XD670 configuration combines eight NVIDIA H100 SXM GPUs on an NVLink HGX baseboard with dual 4th Gen Intel Xeon processors in a 5U air-cooled node. Nexus Compute supplies it configured, integrated, and validated against your fabric and software stack, with HPE warranty through authorized channels.
Who This Solution Is For
Business Benefits
Mature, proven platform
The H100 HGX node is a well-understood building block with broad framework and driver support for predictable rollouts.
Strong training throughput
Eight NVLink-connected H100 GPUs deliver high sustained performance for transformer and vision training at scale.
Air-cooled simplicity
The air-cooled 5U design deploys in standard data centers without facility-water plumbing.
Typical Business Use Cases
Production LLM and NLP model training
Computer-vision and recommendation training
Distributed multi-node deep learning
AI research and experimentation clusters
Industry Applications
Technical Overview
The node uses the NVIDIA HGX H100 8-GPU baseboard with NVLink and NVSwitch, fronted by eight PCIe Gen5 fabric slots for InfiniBand NDR, Slingshot 11, or Ethernet, and driven by dual 4th Gen Intel Xeon Scalable CPUs. Its 5U chassis is air-cooled for standard data-center deployment with up to 32 DDR5 DIMMs.
| GPU / Accelerator | 8x NVIDIA H100 SXM, 80GB HBM3 each (640GB total) |
| GPU Interconnect | NVIDIA NVLink + NVSwitch (HGX baseboard) |
| CPU | Dual 4th Gen Intel Xeon Scalable (Sapphire Rapids) |
| Memory | Up to 32x DDR5-5600 DIMMs |
| Storage | Up to 8x SFF NVMe U.2/U.3 + 2x M.2 RAID |
| Networking / Fabric | 8x PCIe Gen5: InfiniBand NDR / Slingshot 11 / Ethernet |
| Form Factor | 5U single dual-CPU node, air-cooled |
| Management | HPE iLO |
| Warranty | HPE support, configurable terms |
Specifications are indicative and configured to each engagement. Request a quote for a configuration tailored to your requirements.
Warranty, Support & Fulfillment
Every system ships from an authorized channel, configured and tested, with the documentation enterprise buyers need — backed by warranty and a dedicated account team.
Enterprise Warranty
Full manufacturer warranty with optional on-site, next-business-day support and extended coverage.
Authorized Channel
Sourced through Tier-1 distribution and OEM partners — never grey market. Asset & warranty records included.
Lead Time & Deployment
48-hour quotes, then configured, burn-in tested, and delivered on a committed schedule.
Nationwide Fulfillment
Coordinated logistics, rack-and-stack, and delivery wherever your infrastructure lives.
Frequently Asked Questions
Is H100 still a sound choice over H200 or newer GPUs?
Yes for many production pipelines: H100 has the broadest software maturity and strong price-performance, and 640GB aggregate HBM3 handles most training jobs well. Move to H200 only when per-GPU memory capacity is the constraint.
Can this node anchor a multi-node training cluster?
Absolutely. The eight PCIe Gen5 slots accept InfiniBand NDR or Slingshot 11 adapters for low-latency scale-out, and we configure the fabric and topology to match your cluster size.
Does air cooling limit sustained performance?
The air-cooled XD670 is engineered to sustain H100 SXM TDP in a properly provisioned rack. We validate rack power and airflow so performance holds under continuous load.
Hardware Assistance
Configure the HPE Cray XD670 (8x NVIDIA H100 SXM, Dual 4th Gen Intel Xeon) with Nexus Compute
Tell us your requirements and a hardware specialist will help you specify, configure, and quote the right system — typically within two business days. No obligation.