The ML Cloud was established in 2018(?) to aid AI Researchers in their scientific explorations. As such it has been geared towards accelerator cards, suitable for such workloads. The first version of the ML Cloud was entirely Openstack based. Since then, the team has added more traditional HPC components, including job scheduler and hardware expansion.
The ML Cloud consists of two regions: Region 1 with Nvidia 2080ti and V100 accelerator cards, and Region 2 with A100 cards.
Region 3 will be online at the end of 2023 with the new Nvidia generation cards and Weka All-Flash Storage System.
petaFlops
GPUs
per Rack
FULL SYSTEM SUMMARY
This research utilized compute resources at the Tübingen Machine Learning Cloud, DFG FKZ INST 37/1057-1 FUGG.