HPC clusters#
| Cluster name | # nodes | target applications | access | specifics |
|---|---|---|---|---|
| Alex | 82 | high-end GPGPU | NHR, Tier-3 after application | 304 Nvidia A100 (40/80 GB), 352 Nvidia A40 |
| Helma | 96 | high-end GPGPU | BayernKI after application, NHR, Tier-3 (on invitation) | 384 Nvidia H100 (HBM2e 94 GB) |
| Fritz | 1052 | massively parallel, large memory | NHR, Tier-3 after application | parallel filesystem, 64 huge-memory nodes (1 TB, 2 TB RAM) |
| TinyGPU | 35 | GPU | Tier-3 | Nvidia A100, Nvidia RTX3080, Nvidia RTX2080 Ti, Nvidia V100 |
| TinyFat | 47 | large memory | Tier-3 | 256 - 512 GB RAM |
| Woody | 288 | serial throughput | Tier-3 | CPUs with high clock speed |
| Testcluster | ~20 | evaluation of new microarchitectures | request access via hpc-support@fau.de | current Intel/AMD CPUs, AMD/Nvidia GPGPUs, accelerators, early-access prototypes |
Funding#
Tier-3 systems/parts are funded by FAU or as DFG Forschungsgroßgerät while NHR systems/parts are funded by federal and state authorities (BMBF and Bavarian State Ministry of Science and the Arts), respectively. BayernKI is funded by the Free State of Bavaria via its Hightech Agenda Initiative.