Detailed Specifications
HPE P05470-001 NVIDIA Tesla V100 32 GB PCIe Accelerator
Overview
The HPE P05470-001 is a computational GPU accelerator based on NVIDIA's Volta architecture, offering 32 GB of high-bandwidth HBM2 memory. It is designed for AI training, HPC simulations, and data-intensive compute tasks that benefit from massive parallelism. This card fits into a PCIe Gen3 ×16 slot and targets HPE ProLiant and Apollo servers, enabling high-performance workloads within enterprise data centers. With passive cooling and low-profile design, it integrates into chassis environments and leverages HPE's management tools for robust deployment.
Key Features
32 GB HBM2 memory with 900 GB/s bandwidth
5,120 CUDA cores and 640 Tensor Cores for deep learning
Supports FP16, FP32, and FP64 precision for mixed workloads
Passive dual-slot cooling optimized for server racks
PCIe Gen3 ×16 interface for broad server compatibility
Technical Specifications
| Specification | Details |
|---|
| Brand | HPE |
| Part Number | P05470-001 |
| Product Type | GPU Accelerator / Computational GPU |
| GPU Architecture | NVIDIA Volta |
| CUDA Cores | 5,120 |
| Tensor Cores | 640 |
| Memory Capacity | 32 GB HBM2 |
| Memory Bandwidth | 900 GB/s |
| Precision Support | FP16, FP32, FP64 |
| Interface | PCI Express Gen3 ×16 |
| Power Consumption | ~250 W |
| Cooling | Passive, dual-slot |
| Form Factor (H × W × D) | Full-height, dual-slot PCIe |
| ECC Memory | Yes (HBM2) |
| Supported Workloads | AI training, HPC, data analytics |
| Compatible Platforms | HPE ProLiant, Apollo servers |
