Enquire: Asia & Africa - +65-98008081 USA - +1-919-995-4114

Home Networking Inspur AGX-2 AI Supercomputer | NVIDIA Tesla P100

Inspur AGX-2 AI Supercomputer | NVIDIA Tesla P100

Brand:

The Inspur AGX-2 (NF5288M5) is a high-density 2U AI and HPC server designed to combine up to eight NVIDIA Tesla GPUs with dual Intel Xeon processors in a compact rack footprint. This configuration targets deep learning, scientific computing, data analytics, intelligent video, and other GPU-accelerated workloads requiring high compute density, GPU-to-GPU communication, and scalable I/O. The platform supports NVIDIA NVLink 2.0 configurations, high-speed InfiniBand networking, and flexible GPU topologies.

OVERVIEW

  • Inspur AGX-2/NF5288M5 is a high-density 2U GPU computing server for AI, deep learning, HPC, analytics, and accelerated applications, with the requested configuration of 8× Tesla P100 GPUs, 2× Intel Xeon CPUs, and up to 2TB DDR4 memory. Compatibility: designed around supported Intel Xeon, NVIDIA GPU, PCIe/NVLink and InfiniBand configurations. Compare: its key advantage is eight-GPU density in 2U; choose a newer GPU platform when current-generation accelerator support is required. 

USE CASES

  • Deep learning training – Eight Tesla P100 accelerators provide a dense GPU platform for parallel neural-network training workloads.
  • AI model development – Centralizes multiple GPU resources in a compact 2U server for research and development teams.
  • High-performance computing – Supports GPU-accelerated scientific and engineering workloads requiring substantial parallel compute resources.
  • Scientific research – Suitable for computational research, simulation, numerical analysis, and other accelerator-enabled workloads.
  • Data analytics – Provides CPU and GPU resources for computationally intensive analytics pipelines.
  • Intelligent video processing – Supports GPU-accelerated video analysis and computer-vision applications.
  • Machine learning infrastructure – Provides a dedicated GPU compute node for training and inference environments compatible with the platform.
  • HPC cluster deployment – High-speed InfiniBand connectivity and multi-GPU architecture make the platform suitable for cluster-based computing environments. 
  • GPU compute labs – Eight-GPU density helps research organizations consolidate accelerator capacity into fewer rack units.
  • Accelerated enterprise applications – Provides GPU resources for supported applications where CPU-only infrastructure cannot deliver the required parallel processing capacity.

KEY FEATURES

  • 8× NVIDIA Tesla P100 GPU support – Provides high-density GPU acceleration for AI, deep learning, and HPC workloads.
  • 2U high-density design – Packs up to eight GPUs into a compact rack footprint, helping maximize accelerator capacity per rack.
  • Dual Intel Xeon processor architecture – Provides host CPU resources for data preparation, orchestration, application processing, and GPU workloads.
  • Up to 2TB DDR4 memory – Provides substantial system memory capacity for large datasets and memory-intensive applications in supported configurations.
  • NVIDIA NVLink 2.0 support – Enables high-bandwidth GPU interconnect configurations for workloads that require frequent GPU-to-GPU data exchange.
  • Multiple GPU topology options – Allows the platform to be configured for different AI and HPC workload requirements.
  • Up to 16-GPU expansion capability – Supports GPU expansion through a JBOG configuration for scale-up deployments.
  • 100Gbps EDR InfiniBand support – Provides high-speed cluster connectivity for distributed AI and HPC environments.
  • 8× 2.5-inch storage bays – Provides flexible U.2 NVMe, SAS, or SATA storage options for supported configurations.
  • 2× onboard M.2 interfaces – Provides additional local storage options for supported system configurations.
  • Air or air-liquid hybrid cooling – Provides cooling flexibility for high-density accelerator deployments.
  • Dual 3000W Titanium PSUs – Provides high-capacity power delivery for GPU-intensive configurations.
  • Hot-swap storage support – Helps simplify drive maintenance in compatible configurations.
  • IPMI-based management – Enables remote infrastructure management and monitoring through the dedicated management interface. 

TECHNICAL SPECIFICATIONS

  • Brand: Inspur
  • Product: Inspur AGX-2 AI Supercomputer
  • Platform: NF5288M5
  • Form factor: 2U rack server
  • GPU configuration: 8× NVIDIA Tesla P100
  • GPU architecture: NVIDIA Tesla P100
  • GPU interconnect: NVIDIA NVLink 2.0 supported
  • CPU configuration: 2× Intel Xeon Scalable processors
  • Memory: Up to 2TB DDR4-2667 in listed configuration
  • Memory type: DDR4 Registered / LRDIMM support
  • GPU topology: Multiple supported GPU topology configurations
  • GPU expansion: Up to 16 GPUs with supported expansion configuration
  • Storage bays: Up to 8× 2.5-inch U.2/SAS/SATA
  • Onboard storage: 2× M.2 PCIe/SATA
  • Cluster interconnect: Up to 4× 100Gbps EDR InfiniBand HCA support
  • PCIe: PCIe 3.0 expansion support
  • Power supply: 2× 3000W, 80 PLUS Titanium
  • Cooling Redundant: hot-swap fans; air or air-liquid hybrid cooling
  • Management: Dedicated 1000Mbps management interface / IPMI
  • Dimensions: 448 × 87.5 × 899.5 mm
  • Operating temperature: 5°C to 35°C
  • Compatibility: Supported Intel Xeon, NVIDIA GPU, PCIe/NVLink and InfiniBand configurations
  • Intended workloads: AI, deep learning, HPC, scientific computing, analytics, accelerated applications

Why Choose This Product?

  • 8-GPU configuration for high-density AI computing
  • 2U form factor helps conserve valuable rack space
  • Tesla P100 acceleration for supported AI and HPC workloads
  • Dual Intel Xeon processors provide strong host-side processing
  • Up to 2TB DDR4 memory for large data-processing workloads
  • NVLink 2.0 enables high-bandwidth GPU communication
  • 100Gbps EDR InfiniBand supports high-speed cluster networking
  • Dual 3000W Titanium PSUs support demanding GPU configurations
  • Flexible storage with U.2, SAS, SATA and M.2 options
  • Expansion capability supports larger GPU computing environments

CUSTOMER REVIEW

No Reviews Found

FAQ

No FAQs Found