High Performance Computing cluster - supercomputer

Computing environment for scientific and technical tasks. It is implemented as a cluster consisting of head node and several separate worker nodes interconnected with InfiniBand (56 Gbps) technology. The nodes primarily are based on x86 processor architecture but some of them are also equipped with powerful NVidia Tesla graphical accelerators (GPU). HPC cluster can be used in different scientific disciplines from engineering to Big data analytics and machine learning.

Parameters at a glance:

• Total cores: 1200
• Total memory: 8 TB
• Temporary storage volume: 238 TB
• GPU accelerators: 22
• Performance: 139 TFLOPS

Software environment:

• CentOS Linux
• Moab/Torque job management tools
• OpenMPI, MPICH parallel libraries
• Programming: CUDA, Intel C++/FORTRAN, Python, GCC etc.
• Scientific software: MATLAB, COMSOL Multiphysics, GROMACS, Peridigm etc.

• CPU time (core hours)
• GPU time
• High performance virtual desktops
• Parallel software development

• RTU HPC center