HPC facility

LIST's High Performance Computing (HPC) facility mainly supports internal research projects teams in solving advanced and complex computation problems.

The HPC facility consists of a Linux compute cluster, with centralized storage, backup facilities and management infrastructure. The technical specifications of the LIST HPC are the following:

  • 36 disk-less CPU-only compute cluster totaling 1152 CPU cores (Intel Skylake 6130 Gold), 3.4 TB RAM, Intel Omnipath 100 Gbps interconnect, ~ 70 TFlop/s
  • 4 disk-less CPU/GPU nodes with 2 Nvidia V100 GPU per node, totaling 128 CPU cores, 40960 Cuda cores, 5120 Tensor Cores, 768 GB RAM, Intel Omnipath 100 Gbps interconnect, ~ 56 TFlop/s
  • Centralized parallel file system with 321 TB net capacity
  • LTO-7 tape-robot backup system with 256 TB capacity

Below some examples of research that is supported by the HPC (non exhaustive) :

  • Climate modeling simulations
  • Hydrological simulations, flood modeling
  • Molecular dynamics simulations
  • Genome parts matching

Share this page: