Inno3D has introduced a new GPU server built around the NVIDIA MGX architecture, designed to bring modular flexibility and high-density compute to cloud services and AI model deployment. The system targets workloads that demand heavy parallel processing, including high-performance computing, data analysis, machine learning, and deep learning.
New enterprise server targets high-performance computing and AI workloads.
The server is built on Inno3D’s MGX platform and features a sixth-generation Intel Xeon Scalable processor as its core CPU. It uses an eight-channel RDIMM memory configuration with 32 DIMM slots to support high-bandwidth data access for demanding applications.
Spec comparison
| Spec | NVIDIA MGX 4U |
|---|---|
| Form Factor | 4U |
| CPU Platform | Sixth Generation Intel Xeon Scalable Platform |
| Memory Slots | 32 DIMM slots, supporting 8-channel RDIMM |
| GPU Configuration | 8x NVIDIA RTX Pro 6000 Blackwell Server Edition GPUs |
| Power Supply | 3+1 redundant 3200W Titanium PSU |
The primary compute capability comes from eight NVIDIA RTX Pro 6000 Blackwell Server Edition GPUs installed in the chassis. This multi-GPU setup provides significant parallel processing power, making it suitable for rendering, modeling, and AI training tasks that benefit from large-scale GPU clusters.
Power delivery is handled by a 3+1 redundant 3200W Titanium-rated power supply unit. The redundancy ensures system stability during operation, while the high wattage supports the thermal and electrical demands of the eight-GPU configuration.
The server’s design aligns with NVIDIA’s MGX framework for modular edge and data center computing. This architecture allows for flexible deployment in environments where space efficiency and performance density are critical, such as AI inference or cloud-based rendering services.
Discussion
0 comments
Log in to join the thread with a thoughtful take, question, or correction.