GPU Resources

The lab operates two high-performance GPU servers (Microway 5U). The system provides 160 logical CPU cores across two servers with dual processors each. It includes up to 1.5 TB of system memory on the primary machine and approximately 220 TB of total storage, combining local storage with a large shared cluster network drive. This infrastructure supports large-scale data processing and machine learning workloads, enabling efficient training and deployment of advanced AI models.

Server 1 — AMD EPYC System

Microway 5U AMD EPYC "Turin"

ProcessorsDual AMD EPYC 9004/9005 "Turin"

CPU Cores80 logical cores

System Memory1.5 TB

GPUs4× NVIDIA H200 NVL

Total GPU VRAM564 GB

Use CaseHigh-performance training and inference for large-scale machine learning models

Server 2 — Intel Xeon System

Microway 5U Intel Xeon

ProcessorsDual Intel Xeon

CPU Cores80 logical cores

System Memory512 GB

GPUs4× NVIDIA A100 (80 GB) + 3× NVIDIA H100 NVL (95 GB)

Use CaseDeep learning and large language model workloads

Haptic Interaction Devices

The lab utilizes advanced haptic interfaces to support research in perception, sensorimotor interaction, and human–AI behavior.

Force Dimension Omega.3High-precision force-feedback system for studying physical engagement and cognitive responses to force stimuli

Novint Falcon (×4)Tactile feedback controllers for experiments involving motor control, decision-making, and user interaction