Albarqouni Lab · University Hospital Bonn

Lab GPU Server

Our in-house GPU server with 8 NVIDIA A100 80GB GPUs (640 GB GPU memory) for training and evaluating our models.

Apply for access ✉
Illustrative photo of an NVIDIA DGX system (not our server). Photo: Gregory Rocher / Ifremer, CC BY 4.0, via Wikimedia Commons.

Access

Lab members

Hardware

1 node × 8 A100 80GB PCIe · 2 × AMD EPYC 7643 · InfiniBand 200 Gb/s · Slurm

Operated by

Albarqouni Lab — Computational Medical Imaging Research (CoMI)Albarqouni Lab — Computational Medical Imaging Research (CoMI)UKB-IT — IT department of the University Hospital Bonn

The lab’s own GPU server at the University Hospital Bonn, run together with UKB-IT. Its configuration is similar to the A100 nodes of the University’s Marvin and Bender clusters.

  • GPUs — 8 × NVIDIA A100 80GB PCIe (640 GB GPU memory in total), in one node.
  • CPUs — 2 × AMD EPYC 7643, 48 cores each.
  • Network — InfiniBand 200 Gb/s.
  • Software — Linux, Slurm batch system, CUDA.

Access is restricted to members of the Albarqouni Lab with a UKB email account (@ukbonn.de) and an affiliation to our lab. To request an account, use Apply for access above and write from your UKB email address.