Facilities
Center For Artificial Intelligence in Medicine, Imaging & Forensics
NVIDIA DGX H100 Supercomputer & Other Hardware
Specifications
The fourth generation of the most powerful supercomputer for machine learning and AI calculations, equipped with eight Hopper architecture GPUs, delivering unprecedented performance for complex scientific and industrial applications.
| GPU | 8x NVIDIA H100 Tensor Core GPUs |
| GPU memory | 640GB total |
| Performance | 32 petaFLOPS FP8 |
| System power usage | 10.2kW max |
| CPU | Dual Intel Xeon Platinum 8480C (2.0 GHz base, 3.8 GHz boost) |
| System memory | 2TB |
| Networking | 4x OSFP ports, up to 400Gb/s |
| Management network | 2x dual-port ConnectX-7 VPI |
| Storage | OS: 2x 192TB NVMe M.2 |
Other Hardware
Dell R650Xs Servers (2 units)
- 2 Intel Xeon Silver processors (24 cores total)
- 128GB RAM
- 4TB storage
High Performance Virtualization Compute Server
- HexaData 4 GPU 4U Rack Chassis with 2 x 2200W Power Supply
- 2 x Xeon Silver 4210R Processors (10 Cores @ 2.4GHz)
- RAM 256GB, 16TB SSDr
- NVidia Tesla T4 16GB GPU
Server Software
- NVidia Quadro vDWS Subscription, NVidia Virtual Compute Server
- Academic VMware Horizon 7 Enterprise
Software & Tools
BrainVoyager
VERSION : 23.2
Olea
VERSION : 2.3
3D Slicer
VERSION : 5.8.1
AFNI
VERSION : 25.1
FSL
VERSION : 6.0.5
Brainsuite
VERSION : V23A
DSI Studio
VERSION : 2025
Conn (fMRI analysis)
VERSION : V2407
FreeSurfer
VERSION : 7.4.1