EN
Supermicro AS-8125GS-TNHR: 8x H100 SXM5 AI Supercomputer, Dual AMD EPYC Flagship Server for LLM Training & Inference
- Overview
- Related Products
Supermicro AS -8125GS-TNHR | 8U GPU Server
Product Feature
The Supermicro AS-8125GS-TNHR is a flagship GPU supercomputer engineered for ultimate AI computing power. This system is equipped with 8 top-tier NVIDIA H100 SXM5 80GB accelerators, interconnected via a 4th Generation NVLink full-mesh topology, delivering unparalleled surging power for the training, fine-tuning, and real-time inference of trillion-parameter Large Language Models (LLMs). At its core, it features dual AMD EPYC 9004 series processors (such as the EPYC 9754), paired with up to 768GB of DDR5-4800 memory, ensuring robust data orchestration capabilities under ultra-high concurrency. Furthermore, the system is outfitted with 2×100G and 8×400G high-speed network interfaces, completely breaking through the I/O bandwidth bottlenecks in large-scale cluster computing. Whether tackling generative AI, deep learning, or complex scientific simulations, the AS-8125GS-TNHR stands as the ultimate performance choice for enterprises building next-generation AI infrastructure.
Specification
| From Factor | 8U |
| CPU | Dual Socket SP5 |
| Memory | Supports up to 24 DIMM Slots; 4800 MT/s 6TB DDR5 memory |
| Drive Bays | Total of 12 Hot-swap 2.5" NVMe drive bays + 2 hot-swap 2.5" SATA drive bays |
| Power Supply | Total of 6x (3+3) 3000W Redundant Titanium Level Power Supplies |
Supported GPUs |
8 onboard GPUs PCIe 5.0 x16 CPU-to-GPU Interconnect |
Contact Us
Package
Certification
Customer Reviews
