The high-performance DiSVA-HPC computing cluster

The DiSVA-HPC computing cluster is an instrumentation of departmental interest to support all those research activities that require particularly intensive computational capabilities. Thanks to a large number of processors (CPUs), several graphical computing units (GPUs), elevated RAM memory and storage space (HDD), the DISVA-HPC enables analyses ranging from the assembly and comparison of complete genomes to the simulation of evolutionary trajectories of populations and species, from the study of biophysical properties of proteins through molecular dynamics simulations to the implementation of oceanographic models for studying the hydrodynamics of passive tracers. The cluster is also used as a laboratory as part of the teaching activities of the new curriculum in Computational Biology of the Master of Science in Molecular and Applied Biology degree program.

The Disva-HPC cluster currently consists of five Dell/HPE machines configured into a master node with shared filesystem management, login and data storage functions (CPUs = 32; RAM = 128 GB, HDD - raid6 = 32 TB) and four nodes with compute and data storage functions (total: CPUs = 192; RAM = 1TB GB, HDD - raid6 = 20 TB, 3 Tesla T4 GPUs). The nodes are connected to each other via a high-performance infiniband network (50Gb), for sharing home and work areas, and a second Ethernet network (1Gb) for service interactions (ssh, sftp, idrac). The operating system on the machines is a Linux distribution - CentOS7, filesystem sharing is via xCAT, and the work queue management system is based on PBS/Torque.