aoni AI Server
EdgeFrog栖界 NC800
A high-performance compute server for enterprise edge AI inference, serving as the core AI compute device for on-site large model and AI application deployment.
Product Features:
- 13000 TFLOPS (FP8) high-performance AI compute
- 8 x GPU for high-density compute
- 400GE high-speed network for large data transfers
- Air/liquid cooling for high-load scenarios
- PCIe/DDR5/NVMe expansion for various business needs
- Dual redundant power + BMC remote management
| AI Performance | 13000 TFLOPS (FP8) |
| CPU | Up to 2 Intel® Xeon® 4th or 5th Gen Scalable processors; up to 64 cores per CPU, max 350W; supports HBM |
| GPU | GPU × 8 |
| Memory | Up to 32 DDR5 slots; up to 5600MT/s; RDIMM supported; up to 8TB |
| Storage Controller | Integrated 6Gb/s SATA controller; optional 12Gb/s SAS HBA and 12Gb/s SAS RAID controller |
| Storage Expansion | Up to 12×3.5" drives 2×2280/22110 SATA or 1×2280/22110 NVMe M.2 SSD 4 × NVMe SSD |
| PCIe Expansion | Up to 12 standard PCIe slots; up to 10 ×16 standard PCIe slots; 1 x OCP3.0 slot |
| Management | Integrated BMC AST2600 IPMI 2.0, Redfish, SOL, KVM, virtual media |
| Network | 1 x OCP3.0 NIC; optional 1/10/25/50/100/200/400GE PCIe NIC |
| Security | Chassis intrusion detection Intel PFR, SGX TCM/TPM security module |
| Cooling | Air cooling; liquid cooling supported |
| Power Supply | 8×2000W～3200W platinum; N+N redundancy/N+1 hot-plug redundancy |
| Operating Temp | 0℃～35℃ |
| Form Factor | 6U standard chassis |
| Chassis Dimensions (W×H×D) | 263.7 × 447 × 926.9mm |