HPE Cray XD670 Servers with 8x NVIDIA H100 SXM5 80GB GPUs: Special Stock Offer

Aug 12
5 min read

We have HPE Cray XD670 servers with 8x NVIDIA H100 SXM5 80GB HBM3 GPUs on an HGX platform, dual Intel Xeon Platinum 8462Y+ processors, 2TB DDR5 ECC RDIMM memory, ConnectX-7 networking, BlueField-2 DPUs, Intel X710 10G networking, and a 5U form factor in stock, ready to ship.

HPE Cray XD670 NVIDIA H100 Servers

Limited Stock At Special Pricing

Request a Quote

HPE Cray XD670 GPU Server, HPE Cray XD670, Hewlett Packard Enterprise server, HPE GPU server, enterprise AI server, NVIDIA GPU server, GPU-accelerated server, AI infrastructure, high-performance computing, HPC server, AI training server, AI inference server, generative AI server, LLM training server, large language model infrastructure, deep learning server, machine learning server, accelerated computing, enterprise GPU computing, data center AI server, scalable AI infrastructure, HPC infrastructure, enterprise data center server, refurbished HPE Cray XD670, refurbished HPE server, refurbished GPU server, refurbished AI server, refurbished HPC server, refurbished enterprise server, enterprise IT hardware, data center hardware, server-parts.eu

Configuration Overview: HPE Cray XD670 with 8x NVIDIA H100 SXM5 80GB GPUs

The HPE Cray XD670 is designed for organizations that need dense NVIDIA GPU compute, Intel Xeon host processing, large DDR5 memory capacity, and high-speed networking for AI, HPC, research computing, and accelerated data center workloads.

This configuration includes:

HPE Cray XD670 server
5U form factor
8x NVIDIA H100 SXM5 80GB HBM3 GPUs
NVIDIA HGX platform
2x Intel Xeon Platinum 8462Y+ processors
Sapphire Rapids CPU platform
2TB DDR5 ECC RDIMM memory
Memory layout: 32x 32GB
8x ConnectX-7 networking
2x BlueField-2
BlueField-2 connectivity: 2x 25Gb
2x Intel X710 10G networking

The system offers a strong foundation for AI training, large model workloads, HPC, GPU-accelerated analytics, research computing, and enterprise AI infrastructure.

Full Configuration: HPE Cray XD670

Base System: HPE Cray XD670 5U GPU Server

Component	Details	Why it matters
Manufacturer	HPE	Identifies the server vendor
Server Model	Cray XD670	HPE GPU server platform for AI and HPC workloads
Form Factor	5U	Helps buyers plan rack space before deployment
GPU Platform	NVIDIA HGX	Supports dense NVIDIA GPU compute
System Type	GPU server platform	Suitable for AI, HPC, and accelerated data center deployments

The HPE Cray XD670 is suitable for buyers who need a dense NVIDIA H100 GPU server for demanding accelerated compute environments.

CPUs: 2x Intel Xeon Platinum 8462Y+

Component	Details	Why it matters
CPU Model	2x Intel Xeon Platinum 8462Y+	Provides host compute for the GPU platform
CPU Platform	Sapphire Rapids	Identifies the listed Intel Xeon generation
Processor Configuration	Dual CPU configuration	Supports GPU feeding, data handling, and system orchestration
CPU Use Case	AI, HPC, and infrastructure workloads	Supports CPU-side processing around GPU-heavy tasks

This dual-CPU configuration supports workload coordination, data preparation, infrastructure services, and CPU-side processing around the NVIDIA HGX H100 GPU platform.

Memory: 2TB DDR5 ECC RDIMM

Component	Details	Why it matters
Installed Memory	2TB DDR5 ECC RDIMM	Provides large system memory capacity
Memory Layout	64x 32GB	Identifies the installed memory configuration
Memory Type	DDR5 ECC RDIMM	Suitable for enterprise server and HPC workloads
Memory Use Case	AI, HPC, and large dataset workloads	Supports data-heavy workloads alongside GPU acceleration

This memory configuration is suitable for AI pipelines, HPC workloads, research computing, large datasets, and enterprise applications that require large system memory alongside GPU acceleration.

Storage Configuration

Component	Details	Why it matters
Storage Details	Not specified in the provided stock details	Final storage configuration can be confirmed during quotation
Drive Bay Details	Not specified in the provided stock details	Helps avoid incorrect assumptions about storage layout
Storage Use Case	AI datasets, scratch space, and system storage planning	Final storage can be matched to the buyer’s workload requirements

The storage configuration should be confirmed before purchase if the system will be used for local datasets, training pipelines, scratch storage, or workload storage.

PCIe Expansion

Component	Details	Why it matters
GPU Expansion	8x NVIDIA H100 SXM5 GPUs on HGX platform	Provides installed GPU acceleration
Network Expansion	8x ConnectX-7, 2x BlueField-2, 2x Intel X710 10G	Supports high-speed networking and data center connectivity
Expansion Use Case	GPU acceleration, networking, and HPC infrastructure	Useful for AI, HPC, and research computing deployments

The installed GPU and networking configuration supports environments where acceleration, high-speed data movement, and data center integration are important.

Power & Cooling

Component	Details	Why it matters
Power Supply Details	Not specified in the provided stock details	Final PSU configuration can be confirmed during quotation
Power Rating	To be confirmed	Important for GPU server deployment planning
Cooling Configuration	Not specified in the provided stock details	Final cooling details can be confirmed during quotation
Thermal Planning	Required for 5U 8-GPU server deployment	Helps buyers plan rack, airflow, and data center requirements

The power and cooling configuration should be confirmed during quotation so the final deployment can be planned correctly for a dense 8-GPU HPE Cray XD670 server.

Platform Overview: HPE Cray XD670

The HPE Cray XD670 is designed for organizations that need:

8x NVIDIA H100 SXM5 GPU acceleration
80GB HBM3 memory per GPU
640GB total GPU memory
NVIDIA HGX platform support
Dual Intel Xeon Platinum host processors
2TB DDR5 ECC RDIMM system memory
High-speed ConnectX-7 networking
BlueField-2 DPU support
Intel X710 10G networking
5U GPU server deployment
AI, HPC, and research infrastructure support

This server platform is a good fit for AI infrastructure teams, research groups, data center buyers, and enterprise IT teams looking to deploy NVIDIA H100 GPU systems for accelerated computing.

Use Cases: HPE Cray XD670 with 8x NVIDIA H100 SXM5 GPUs

AI model training
Large language model workloads
Deep learning infrastructure
High-performance computing
GPU-accelerated analytics
Research computing
Scientific simulation
AI inference workloads
Distributed training environments
Data center GPU infrastructure
Multi-GPU workload development
High-speed network-connected AI workloads

The HPE Cray XD670 configuration makes this server useful for organizations that need dense NVIDIA H100 GPU compute, large system memory, Intel Xeon host processing, and high-speed networking in a 5U platform.

Testing, Condition, and Warranty: HPE Cray XD670

Condition

Stock server configuration
Hardware details based on available configuration information
Final power, cooling, PCIe, and storage layout details can be confirmed during quotation

Testing Includes

GPU platform inspection
CPU, memory, and system checks
ConnectX-7 adapter inspection
BlueField-2 inspection
Intel X710 10G networking inspection
Storage configuration inspection where applicable
Power and cooling configuration inspection where applicable
Basic server hardware validation

Warranty

Warranty information available on request
Additional support options may be available depending on stock and configuration
Final warranty terms confirmed during quotation

FAQ: HPE Cray XD670 NVIDIA H100 Server

What is the HPE Cray XD670 with NVIDIA H100 used for?

The HPE Cray XD670 with NVIDIA H100 GPUs is used for AI training, large language models, deep learning, HPC, GPU-accelerated analytics, research computing, scientific simulation, and enterprise AI infrastructure.

How many GPUs are included in this HPE Cray XD670 configuration?

This HPE Cray XD670 configuration includes 8x NVIDIA H100 SXM5 80GB HBM3 GPUs on an NVIDIA HGX platform, providing 640GB total GPU memory.

What CPUs are included in this HPE Cray XD670 server?

This configuration includes 2x Intel Xeon Platinum 8462Y+ processors on the Sapphire Rapids platform. These CPUs provide host compute for the NVIDIA HGX H100 GPU system.

What networking is included in this HPE Cray XD670?

This HPE Cray XD670 includes 8x ConnectX-7, 2x BlueField-2 with 2x 25Gb connectivity, and 2x Intel X710 10G networking.

Where can I buy HPE Cray XD670 NVIDIA H100 servers in Europe?

You can request a quote from server-parts.eu for HPE Cray XD670 servers, NVIDIA H100 SXM5 GPU servers, HGX H100 8-GPU systems, AI servers, HPC servers, and related data center hardware.