About the system
SuperMUC-NG consists of Intel Xeon 'Skylake' processors
In total there are 311,040 compute cores with a main memory of 719 TB and a peak performance of 26.9 PetaFlop/s.
In addition to the compute nodes there are 64 Cloud Nodes (half of them equipped with 2 GPUs each).
The internal interconnect is a fast OmniPath network with 100 Gbit/s.
The compute nodes are bundled into 8 domains (islands).
Using the System
- Access and Login to SuperMUC-NG
- Application for a project on SuperMUC-NG
- Data Migration from SuperMUC to SuperMUC-NG
- File Systems of SuperMUC-NG
- Hardware of SuperMUC-NG
- Job Processing with SLURM on SuperMUC-NG
- Operational Concept
- Usage Statistics for SuperMUC
- User Affairs of SuperMUC-NG
- Building and Running applications on SuperMUC-NG
- Compute Cloud of SuperMUC-NG
- SuperMUC-NG Status
- Software for HPC (see also: lrztools and lrzlib on SuperMUC-NG, or command:
- LRZ Compute Cloud
- HPC Calls for projects, allocations, support, or funding
- Courses, Training and Events for HPC
- Public Relations for HPC (including the scientific results obtained)