An HPC Cluster is a group of individual computer systems interconnected by a shared network for the purposes of carrying out High Performance Computing and High Throughput Computing.
The main features of our HPC Clusters are:
- Multiple compute ‘nodes of relatively low cost hardware
- x86_64 CPU architecture
- CentOS Linux v7
- Gbit/10Gbit ethernet interconnect
- Common storage visible to all compute nodes
- Workload scheduler (SLURM)
- Multiple job execution queues
- Shared software repositories (> 200 bioinformatics codes and libraries)
- Virtualised login nodes for user access and job submission
We currently manage a cluster estate of ca 3500 cores, directly connected to our EMC Isilon shared storage system.