Safe checkout
Secure payments
Fast dispatch
1–2 day handling
Free shipping
U.S. ground
Warranty & RMA
Support after delivery
HPE P24251-001 1 InfiniBand MCX653105A-ECAT PCIe4 x16 Network Adapter
- Dual-port InfiniBand HDR100 connectivity.
- Based on Mellanox ConnectX-6 MCX653105A-ECAT technology.
- Utilizes QSFP56 connectors for high-speed links.
- PCIe Gen4 x16 host interface for maximum bandwidth.
- Supports up to 200Gbps aggregate bandwidth (100Gbps per port).
- Optimized for High-Performance Computing (HPC) and Artificial Intelligence (AI) / Machine Learning (ML) applications.
- Low latency and high message rate capabilities.
- Includes advanced features like RDMA and GPUDirect.
Was $680.00 · You save $0.00
Key specifications
See full specs ↓- Product Type
- Network Adapter
- Technology
- InfiniBand HDR100
- Chipset
- Mellanox ConnectX-6 MCX653105A-ECAT
- Port Configuration
- 2 x QSFP56
- Speed
- HDR100 (100Gbps per port)
- Interface
- PCIe 4.0 x16
Product details
The HPE P24251-001 is a high-performance InfiniBand network adapter featuring the Mellanox ConnectX-6 MCX653105A-ECAT chipset. It offers dual-port connectivity with HDR100 speeds over a PCIe Gen4 x16 interface, designed for demanding HPC and AI/ML workloads.
Technical Information
| Product Type | Network Adapter |
| Technology | InfiniBand HDR100 |
| Chipset | Mellanox ConnectX-6 MCX653105A-ECAT |
| Port Configuration | 2 x QSFP56 |
| Speed | HDR100 (100Gbps per port) |
Additional Specifications
| Interface | PCIe 4.0 x16 |
| Form Factor | Standard PCIe Card |
| Manufacturer | HPE |
| Part Number | P24251-001 |
Description
The HPE P24251-001 is a cutting-edge InfiniBand adapter designed to deliver exceptional performance for the most demanding computational tasks. Leveraging the power of the Mellanox ConnectX-6 MCX653105A-ECAT controller, this adapter provides two ports, each capable of operating at InfiniBand HDR100 speeds (100Gbps), resulting in a total aggregate bandwidth of up to 200Gbps. The adapter utilizes a PCIe Gen4 x16 interface, offering significantly higher bandwidth and lower latency compared to previous generations. This makes it an ideal solution for tightly coupled clusters used in High-Performance Computing (HPC), large-scale data analytics, and the training of complex Artificial Intelligence and Machine Learning models, where rapid data exchange is critical. With advanced features such as Remote Direct Memory Access (RDMA) and GPUDirect, the P24251-001 minimizes CPU overhead and accelerates data transfers directly between memory and devices, including GPUs. This optimization is crucial for achieving maximum application performance and efficiency in modern supercomputing and AI infrastructure.
Condition: Refurbished
Free U.S. ground shipping
Typically 1–2 days handling, 3–7 days in transit.
Purchase orders accepted
Government, enterprise, data center and small business.
Buying 10+ units?
Get volume pricing and availability for P24251-001.
