High Performance Computing

This private cloud service runs on a homogenous cluster containing 250+ physical servers with 7000+ physical cores, 90+ TB RAM and Petabytes of usable storage space. Powered by Intel Xeon Gold Scalable processors with AVX-512 instruction set supporting up to 2x FLOPS per core per cycle. Performance boost options using add-in datacenter GPUs.

Service Contains

Preinstalled & preconfigured modules

  • Solvers (e.g. OpenFOAM for CFD), compilers, visualization, utilities, etc.
  • Both in Debug & Release configurations
  • With configured MPI and/or OpenMP/TBB parallel options

Fully setup environment with usage instructions for each component

  • Modules, disks, libraries, configuration files

Managed execution

  • Job scheduling, job monitoring, e-mail notifications, usage statistics
  • Simple selection to run a job on a single machine, logical cluster part,
    entire physical cluster

Managed access

  • SSH configuration, common & private disk space, disk quotas, fast results file transfer

Platform

  • Parallelism-supporting libraries and platforms (MPI, OpenMP, Intel TBB – license may be required)
  • High-performance computing libraries (Intel MKL, BLAS, LAPACK, ScaLAPACK, Arpack, PETSc, SuiteSparse, etc.) set up