Wholesale brand new SYS 422GA NBRT LCC SYS422 GL NR Complete liquid-cooled rack integration is required for purchase Dual
Here’s a professional comparison table in the style of Supermicro’s documentation, comparing SYS-422GA-NBRT-LCC and SYS-422GL-NR:
Supermicro 4U GPU Server Comparison
| Feature | SYS-422GA-NBRT-LCC (4U Liquid Cooled) | SYS-422GL-NR (4U Air Cooled) |
|---|---|---|
| Form Factor | 4U Rackmount with LCS | 4U Rackmount |
| CPU Support | Dual 4th Gen Intel® Xeon® SP (350W) | Dual 4th Gen Intel® Xeon® SP (350W) |
| GPU Capacity | 8x NVIDIA SXM5 GPUs (H100/H200) | 10x PCIe GPUs (L40S/A100) |
| GPU Cooling | Direct-to-Chip Liquid Cooling | Optimized Airflow Tunnel |
| Memory | 32x DDR5 DIMMs (8TB max) | 24x DDR5 DIMMs (6TB max) |
| Storage | 12x 2.5″ NVMe (Front) + 4x 2.5″ (Rear) | 24x 2.5″ NVMe/SAS/SATA |
| PCIe Expansion | 8x PCIe 5.0 x16 (GPU-optimized) | 10x PCIe 5.0 x16 |
| Networking | 4x 200G ConnectX-7 + OCP 3.0 | 2x 100G ConnectX-6 + OCP 3.0 |
| Power Supply | 5600W Titanium (N+N) | 3200W Platinum (N+1) |
| Cooling System | Full liquid cooling (CPU+GPU+PCIe) | Advanced Air Cooling |
| Rack Density | 44kW per rack | 28kW per rack |
| Ideal Workloads | Exascale AI Training, HPC | Enterprise AI, Rendering |
Here are detailed descriptions for both Supermicro 4U GPU servers:
Supermicro SYS-422GA-NBRT-LCC – 4U Liquid-Cooled AI Supercomputing Server
The Ultimate in Data Center Efficiency
The SYS-422GA-NBRT-LCC represents Supermicro’s most advanced liquid-cooled solution for extreme-density AI workloads. This 4U rackmount system combines direct-to-chip liquid cooling with 8x NVIDIA HGX H100/H200 SXM5 GPUs, delivering unparalleled performance for large language model training and HPC applications.
Key Features:
-
Breakthrough Cooling Technology:
-
Full-immersion liquid cooling for CPUs, GPUs, and PCIe devices
-
Supports 44kW per rack density (4x traditional air-cooled)
-
95% heat removal via liquid for ultra-efficient operation
-
-
Unmatched GPU Performance:
-
8x NVIDIA SXM5 GPUs (H100/H200) with full NVLink connectivity
-
700W TDP per GPU with optimized thermal transfer
-
NVIDIA HGX baseboard with 900GB/s GPU-to-GPU bandwidth
-
-
Enterprise-Class Infrastructure:
-
Dual 4th Gen Intel® Xeon® Scalable (350W TDP)
-
32x DDR5 DIMM slots (8TB memory)
-
12x front + 4x rear NVMe drives (PCIe 5.0)
-
-
Hyperscale Networking:
-
4x 200G NVIDIA ConnectX-7 SmartNICs
-
OCP 3.0 mezzanine slot
-
BlueField-3 DPU ready
-
Ideal For: National labs, hyperscale AI training, quantum computing research
Supermicro SYS-422GL-NR – 4U Universal GPU Workhorse
Maximum Flexibility for Enterprise AI
The SYS-422GL-NR delivers exceptional air-cooled GPU density in a standard 4U form factor, supporting up to 10x full-length PCIe GPUs for diverse accelerated computing needs.
Key Features:
-
Unprecedented GPU Flexibility:
-
Supports 10x 350W PCIe GPUs (L40S, A100, H100 PCIe)
-
Supermicro’s GPU Air Shroud technology
-
Tool-less GPU maintenance design
-
-
Massive Storage Capacity:
-
24x 2.5″ hot-swap bays (all NVMe capable)
-
Hardware RAID options
-
Dual M.2 boot drives
-
-
Balanced Performance:
-
Dual 4th Gen Intel® Xeon® Scalable
-
24x DDR5 DIMM slots (6TB)
-
10x PCIe 5.0 x16 slots
-
-
Enterprise-Ready:
-
3200W Platinum PSUs (N+1)
-
Standard 42°C ambient operation
-
Front-accessible service panels
-
Ideal For: Enterprise AI, render farms, financial modeling, healthcare imaging
