HPC
In July 2016 we installed a high performance computing system (HPC). This cluster is specifically designed for pre-processing of next generation sequencing (NGS) data. However, any application that requires massive parallel processing could be implemented as well.
A short introductory documentation can be found here.
Hardware
The HPC consists of
- 33 servers (nodes) that can perform calculations,
- file storage system with an effective size of 110 TB (BeeGFS file system)
- fast interconnect (Infiniband FDR) that allows for fast file transfer and MPI applications
The following scheme illustrates the structure of the cluster.
Software
Operating system: CentOS 7
xCAT (Extreme Cloud/Cluster Administration Toolkit)
SLURM (job scheduling system)
Access
All member of the BMC can get access. Please contact us for details