AS-8126GS-NB3RT
DP AMD 8U GPU System with NVIDIA HGX B300 NVL8
- Dual AMD EPYC™ 9005/9004 Series Processors
- 8x Blackwell 300 Ultra SXM GPUs
- Supports up to 24 DIMM slots, 6400 MT/s 6TB DDR5 memory (1DPC only)
(Please check System Memory Section for detail) - 8x OSFP 800GbE
- 2x PCIe 5.0 x16 FHHL slots
- 8x front hot-swap E1.S NVMe drive bays
- 6x 6600W Redundant (3 + 3) Titanium Level (96%) power supplies
(Power supply full redundancy based on configuration and application load)
Key Applications
- Scientific Research
- Conversational AI
- Business Intelligence & Analytics
- Drug Discovery
- Finance Services and Fraud Detection
- AI/Deep Learning Training and Inference
- Large Language Model (LLM) and Generative AI
- High Performance Computing (HPC)
- Autonomous Vehicle Technologies
Datasheet
Product Specification
| Product SKUs | A+ Server AS -8126GS-NB3RT |
| Motherboard | Super H14DSG-OM |
| Processor | |
| CPU | Dual processor(s) AMD EPYC™ 9005/9004 Series Processors |
| Core Count | Up to 384C/768T |
| Note | Supports up to 500W TDP CPUs |
| GPU | |
| Max GPU Count | 8 onboard GPUs |
| Supported GPU | NVIDIA SXM: HGX B300 8-GPU (288GB) |
| GPU-GPU Interconnect | NVIDIA® NVLink® with NVSwitch™ |
| System Memory | |
| Memory | Slot Count: 24 DIMM slots Max Memory (1DPC): Up to 6TB 6400MT/s ECC DDR5 RDIMM (AMD EPYC™ 9005 Series Processor) Max Memory (1DPC): Up to 6TB 4800MT/s ECC DDR5 RDIMM (AMD EPYC™ 9004 Series Processor) |
| Memory Voltage | 1.1V |
| On-Board Devices | |
| Chipset | System on Chip |
| Network Connectivity | 8 OSFP 800GbE |
| Input / Output | |
| USB | 1 USB 3.0 Type-A port(Front) 1 USB 2.0 Type-A port(Front) |
| Video | 1 VGA port 1 mini-DP port |
| System BIOS | |
| BIOS Type | AMI 64MB UEFI |
| Management | |
| Software | SuperCloud Composer® Supermicro Server Manager (SSM) Super Diagnostics Offline (SDO) Supermicro Thin-Agent Service (TAS) SuperServer Automation Assistant (SAA) New! |
| Power configurations | Power-on mode for AC power recovery ACPI Power Management |
| Security | |
| Hardware | Trusted Platform Module (TPM) 2.0 Silicon Root of Trust (RoT) – NIST 800-193 Compliant |
| Features | Cryptographically Signed Firmware Secure Boot Secure Firmware Updates Automatic Firmware Recovery Supply Chain Security: Remote Attestation Runtime BMC Protections System Lockdown |
| PC Health Monitoring | |
| CPU | Monitors for CPU Cores, Chipset Voltages, Memory 8+4 Phase-switching voltage regulator |
| FAN | Fans with tachometer monitoring Status monitor for speed control Pulse Width Modulated (PWM) fan connectors |
| Temperature | Monitoring for CPU and chassis environment Thermal Control for fan connectors |
| Chassis | |
| Form Factor | 8U Rackmount |
| Model | CSE-GP807TS-R000NP |
| Dimensions and Weight | |
| Height | 13.8" (356 mm) |
| Width | 17.6" (449 mm) |
| Depth | 37.4" (950 mm) |
| Package | 27.6" (H) x 51" (W) x 29.5" (D) |
| Weight | Gross Weight: 266.76 lbs (121 kg) Net Weight: 108 lbs (49.4 kg) |
| Available Color | Silver |
| Front Panel | |
| LED | HDD activity LAN1 activity LAN2 activity Power Fail Power status System information (overheat/UID) |
| Buttons | Power On/Off UID |
| Expansion Slots | |
| PCI-Express (PCIe) Configuration | Default 2 PCIe 5.0 x16 FHHL slots |
| Drive Bays / Storage | |
| Drive Bays Configuration | Default: Total 8 bays 8 front hot-swap E1.S NVMe drive bays |
| M.2 | 2 M.2 NVMe slots (M-key) |
| System Cooling | |
| Fans | Up to 12 heavy duty fans with optimal fan speed control |
| Operating Environment | |
| Environmental Spec. | Operating Temperature: 10°C to 35°C (50°F to 95°F) Non-operating Temperature: -40°C to 60°C (-40°F to 140°F) Operating Relative Humidity: 8% to 90% (non-condensing) Non-operating Relative Humidity: 5% to 95% (non-condensing) |
Air-Cooled Supermicro NVIDIA HGX™ B300 8-GPU Servers
The Proven Platform for Large-Scale AI Workloads, Upgraded with Ultra Performance
The Supermicro NVIDIA HGX platform is the building block of the world’s largest AI training clusters, delivering the immense computational output required for delivering the transformative AI applications of today. Now featuring NVIDIA Blackwell Ultra, Supermicro’s NVIDIA HGX B300 solutions enable rapid deployment of the industry’s highest-performing AI infrastructure to tackle the largest-scale AI training, real-time AI reasoning, agentic AI applications, multimodal AI inference, and physical AI applications.
Supermicro integrates Blackwell Ultra generational advancements with system-level and rack-level engineering to create a solution that maximizes efficiency and performance at scale. Both 8U air-cooled and 4U liquid-cooled configurations are available, providing up to 7.5x performance gains over NVIDIA Hopper systems, along with 288GB of HBM3e memory per GPU* to tackle massive AI model sizes.
NVIDIA HGX™ B300 8-GPU Server
SYS-822GS-NB3RT / AS -8126GS-NB3RT
| Overview | 8U air-cooled system with front I/O NICs, DPUs, storage, and management |
|---|---|
| CPU |
|
| Memory |
|
| GPU | NVIDIA HGX B300 8-GPU (288GB HBM3e per GPU*) 1.8TB/s NVLink GPU-GPU interconnect with NVSwitch |
| Networking | 8 integrated NVIDIA ConnectX®-8 SuperNICs, up to 800Gb/s 2 dual-port NVIDIA BlueField®-3 DPUs |
| Storage | 8 Hot-swap E1.S NVMe storage drive bays and 2 M.2 NVMe boot drives |
| Power Supply | 6x 6600W Redundant (3 + 3) Titanium Level (96%) power supplies |
