Technical specifications of the High Performance Computing (HPC) environment on Gearshift
Software
Key ingredients of the High Performance Computing (HPC) environment of the Gearshift cluster
- Linux OS: CentOS 7.9 with Pulp for package distribution/management.
- Job scheduling: Slurm Workload Manager 22.05.2
- Module system: Lmod
- Deployment of (Bioinformatics) software: EasyBuild
Virtual Servers
- Jumphosts: airlock
- User Interfaces (UIs): gearshift
- Deploy Admin Interfaces (DAIs): sugarsnax
- Sys Admin Interfaces (SAIs): imperator
- Compute Nodes: gs-vcompute01, gs-vcompute02, gs-vcompute03, gs-vcompute04, gs-vcompute05, gs-vcompute06, gs-vcompute07, gs-vcompute08, gs-vcompute09, gs-vcompute10
Shared Storage
A Logical File System (LFS) is usually a piece of a larger Physical File System (PFS) that serves a specific need for a specific user group. In case it as a network file system you could call it a share. In addition to LFS-ses for home dirs and the centrally deployed software and reference data the Gearshift HPC cluster has access to the following LFS-ses:
- Available prm LFS-ses: prm01, prm02, prm03
- Available rsc LFS-ses: rsc01
- Available tmp LFS-ses: tmp01
Resources available to Slurm jobs
regular partition
| Resource | Amount/value |
|---|---|
| Number of nodes | 10 |
| Cores/node | 22 |
| RAM/node (MB) | 205490 |
| Storage/node (MB) | 2800000 |
| Node features | rsc01,tmp01 |
user_interface partition
| Resource | Amount/value |
|---|---|
| Number of nodes | 1 |
| Cores/node | 1 |
| RAM/node (MB) | 1024 |
| Storage/node (MB) | 0 |
| Node features | prm02,prm03,tmp01 |