| Customization: | Available |
|---|---|
| Application: | Enterprise Level |
| Instruction System: | Cuda Ptx |
Suppliers with verified business licenses
Audited by an independent third-party inspection agency
Core architecture | Ampere Architecture (Quad-GPU Design) |
Memory configuration | 4×16GB GDDR6 with ECC (Error Correcting Code) |
Video memory bandwidth | 4×200 GB/s (Total 800 GB/s) |
CUDA core | 4×1280 CUDA Cores (Total 5120) |
Tensor Core | 3rd-gen Tensor Cores (4×40) |
RT Core | 2nd-gen RT Cores (4×10) |
Computational performance | FP32: 4×4.5 TFLOPS FP16: 4×17.9 TFLOPS INT8: 4×35.9 TOPS |
Interface and bus | PCIe 4.0 ×16 (PCIe Gen4 Compatible) |
Power consumption and heat dissipation | 250W TDP, Passive Cooling (2-slot FHFL) |
Video codec | 4×NVENC (Encoder) / 8×NVDEC (Decoder) with AV1 Decode Support |
Virtualization support | NVIDIA vPC/vWS/vApp/vCS, Up to 64 Concurrent Users |
Physical size | Full-Height Full-Length (FHFL), Dual-slot Form Factor |
Safety characteristics | Hardware Root of Trust (Secure Boot) |
Applicable scenario | Graphics-intensive VDI, AI Inference, Multi-user Workstation Rendering |





