Inspur NF5688M7 BareBone-Intel Xeon Platinum 8480+, 1,5 TB DDR5, nó HGX-H100 GPU
  • Categorias de Produtos:Servidores
  • Número da peça:Inspur NF5688M7
  • Disponibilidade:In Stock
  • Condição:Novo
  • Características do produto:Pronto para envio
  • Pedido mínimo:1 unidade
  • Preço da lista foi:$347,567.00
  • Seu preço: $317,428.00 Você economiza $30,139.00
  • Chat agora Enviar email

Fique tranquilo. Devoluções aceitas.

Frete: O envio internacional de itens pode estar sujeito a processamento aduaneiro e encargos adicionais. Ver detalhes

Entrega: Por favor, permita mais tempo se a entrega internacional estiver sujeita ao processamento aduaneiro. Ver detalhes

Devoluções: Devoluções em 14 dias. O vendedor paga pelo frete de devolução. Ver detalhes

Frete GRÁTIS. Estamos aceitando pedidos de compra NET 30 Dias. Obtenha uma decisão em segundos, sem afetar seu crédito.

Se você precisar de uma grande quantidade do produto Inspur NF5688M7 - ligue para nós pelo nosso número gratuito Whatsapp: (+86) 151-0113-5020 ou solicite um orçamento no chat ao vivo e nosso gerente de vendas entrará em contato com você em breve.

Inspur NF5688M7 Barebone – Dual Intel Xeon Platinum 8480+, 1.5 TB DDR5, HGX-H100 GPU Node

Keywords
Inspur NF5688M7, Xeon Platinum 8480+, DDR5 4800MT/s, HGX-H100 8-GPU, U.2 NVMe storage, GPU compute server

Description

The Inspur NF5688M7 is a 6U high-density barebone server engineered for AI, HPC, and large-scale compute workloads. Featuring dual Intel Xeon Platinum 8480+ CPUs—each delivering 56 cores and 112 threads—it provides unmatched multi-threaded processing power.

Armed with 24 × 64 GB DDR5-4800 MT/s RDIMMs, the system boasts an impressive 1.5 TB of ultra-fast memory bandwidth, ideal for shifting massive datasets and AI model parameters.

Graphic acceleration is driven by an NVIDIA HGX-H100 module housing eight H100 GPUs, leveraging the innovative Hopper architecture for acceleration across AI and machine learning workloads.

For storage, the system includes two 1.9 TB M.2 NVMe SSDs for system operations and four 3.84 TB U.2 NVMe SSDs for ultra-low latency data needs, enabling rapid data retrieval and processing.

Connectivity is top-tier: a dual-port NVIDIA ConnectX-7 PCIe Gen4 100 Gb NIC and a 400 Gb ConnectX-7 PCIe Gen5 1-port NIC deliver peak network throughput for AI clusters and disaggregated workflows.

Power redundancy is designed for mission-critical uptime—two 3200 W and six 2700 W platinum-level PSUs guarantee uninterrupted operation even in adversarial power conditions. This package includes rack rails and comes with a 3-year U.S. standard RMA warranty.

Key Features
  • Dual Intel Xeon Platinum 8480+ (56C/112T each) for parallel compute efficiency
  • 1.5 TB DDR5-4800 MT/s high-bandwidth memory
  • 8-GPU NVIDIA HGX-H100 module delivering industry-leading AI performance
  • Mixed ultra-fast storage: 2×1.9 TB M.2 NVMe + 4×3.84 TB U.2 NVMe
  • High-speed networking: dual 100 Gb PCIe Gen4 and single 400 Gb PCIe Gen5 NICs
  • Robust power architecture: 2×3200 W + 6×2700 W platinum PSUs
  • 3-year standard on-site RMA in the USA for enterprise peace of mind
Configuration
ComponentSpecification
Form Factor6U Rack Module
Processors2× Intel Xeon Platinum 8480+ (56C/112T)
Memory24× 64 GB DDR5-4800 MT/s (1.5 TB)
GPU Module1× NVIDIA HGX-H100 with 8 GPUs
System Storage2× 1.9 TB M.2 NVMe SSDs
Data Storage4× 3.84 TB U.2 NVMe SSDs
Network Interfaces1× 100 Gb dual-port PCIe Gen4, 1× 400 Gb single-port PCIe Gen5 (ConnectX-7)
Power Supplies2× 3200 W + 6× 2700 W Platinum redundant
AccessoriesRack rails, USA power cables
Warranty3-year standard RMA in USA
Compatibility
This server is validated for NVIDIA HGX-H100 GPU modules and Intel Xeon Platinum 8480+ processors, aligned with high-performance DDR5 platforms and ConnectX-7 InfiniBand/Ethernet for AI clusters.

Usage Scenarios

Optimized for large-scale AI training, including language models, generative AI, and Transformer architectures.

Ideal for HPC workflows needing vast compute, memory, and storage bandwidth.

Tailored for multi-node AI clusters leveraging high-bandwidth networking and GPU interconnectivity.

Ready for enterprise data centers demanding high availability, scalable performance, and easy integration into GPU-based compute fabric.

Frequently Asked Questions
Q1: What performance benefits does the HGX-H100 GPU bring?**
With 8 H100 SXM GPUs, the module delivers up to 32 PFLOPS of FP8 performance and 640 GB of HBM3 memory.

Q2: Why use DDR5-4800 memory?**
DDR5-4800 provides superior bandwidth and capacity, vital for feeding GPUs and CPUs in AI workloads efficiently.

Q3: What networking speeds are supported?**
The system offers dual 100 Gb PCIe Gen4 and a 400 Gb PCIe Gen5 NIC, excellent for ultra-fast inter-node connectivity, such as InfiniBand fabrics.

Q4: How does the power configuration support reliability?**
Redundant platinum-level PSUs guarantee full uptime and hot-swap capability during failure or maintenance.

PRODUTOS RELACIONADOS A ESTE ITEM
HPE Proliant DL380 Gen11 Rack Server - Xeon Gold5416s, 128 GB DDR5, 6 × 3,84 TB SSD, MR408I -OR Recomendar
Lenovo Thinksystem SR650-Dual Intel Xeon Gold6230 Servidor Dual-CPU com RAM de 256 GB e 8 × 1,2 TB 10K SAS Storage Recomendar
Lenovo Thinksystem SR860 Server - Dual Intel Xeon Gold 5118 CPUs, múltiplas configurações para aplicativos de ERP, virtualização e banco de dados Recomendar
Lenovo Thinksystem SR850 Server, 2x Intel Xeon Gold 5120, 64 GB de RAM, servidor de rack de quatro vias 2U, Bays de disco de 2,5 " Recomendar
Dell EMC PowerEdge R760 Chassis, Baías de 24 x 2,5 ", fontes de alimentação de alcance quente de 2 x 1400W Recomendar
Lenovo Thinksystem SR850 Servidor - Quad Intel Xeon Platinum 8280, RAM 24x64 GB, armazenamento de 2x300 GB, Gigabit NIC de 4 portas, fonte de alimentação, software de gerenciamento, nova embalagem Recomendar
Servidor Lenovo SR670v2, duplo 6334 8-core 3,6 GHz, 4x32GB RAM, 4x960GB SSD, 730-8i RAID, fonte de alimentação dupla 2000W Recomendar
Dell PowerEdge R840 - 24 discos, sem CPU, 2 GPUs discretas, sem memória, 3x 600 GB 10K HDD, H740P, PSUS de 1600W dual, kit de trilho, serviço de 3 anos Recomendar