Aiserver Manufacturer
1 / 5
| Product Model | xi Si N260 (MXN260) 64GB |
| Market Positioning | Domestic high-end training and inference integrated GPU, optimized for AI large model training/inference and scientific computing. Comparable to L20. |
| Computing Power (FP32) | 15 TFLOPS |
| Computing Power (FP16) | 80 TFLOPS |
| Computing Power (INT8) | 160 TOPS |
| Memory Type | HBM2E (High Bandwidth Memory, approx. 30% higher bandwidth than HBM2) |
| Process Technology | TSMC 7nm FinFET process, paired with 2.5D CoWoS advanced packaging |
| Memory Bandwidth | 1.2TB/s, complying with HBM2 specification characteristics. |
| Memory Bus Width | 1024bit |
| Effective Frequency | 9.375Gbps (Equivalent frequency 4000 MHz) |
| Latency | ~7ns (Physical characteristics of HBM2E) |
| Interface | PCIe Gen4.0 x16 (Bidirectional bandwidth 64GB/s), supporting x86/ARM host interconnection. |
| Form Factor | FHFL dual-slot, Weight: 1064g |
| Graphics Support | Does not support DirectX or OpenGL; focused on data center inference/computing scenarios. |
Cyber Resilient Architecture
Silicon-based root of trust anchors end-to-end boot resilience while Multi-Factor Authentication (MFA) and role-based access controls ensure trusted operations.
Autonomous Collaboration
Simplify, automate and centralize one-to-many management with the Enterprise console and iDRAC solutions for enterprise servers.
Sustainability
Innovative options for energy efficiency designed to reduce the carbon footprint and lower operation costs through thoughtful engineering.
Designed for 24x7 enterprise data center operations, the xi Si N260 (MXN260) features hardware optimized for deployment at scale. It supports Secure Boot with Hardware Root of Trust technology, providing an extra layer of security and meeting NEBS Level 3 standards.
The passively cooled, energy-efficient design fits a wide range of certified systems from OEM vendors and is easy to integrate into existing data center infrastructure.
