The HPC Laboratory involves the use of cluster-type parallel computing systems and has been operating at USM since 2006 within the Center for Education and Research in Mathematics and Informatics (CECMI).

The purpose of the HPC Laboratory is to implement new HPC technologies in the training and research process, for the development of parallel algorithms and their execution on local cluster-type systems and even on regional Grid or Cloud-type systems.

The available HPC technical base allows for parallel programming in high-performance UNIX environments, the creation and management of large databases (electronic libraries, electronic courses, etc.), the provision of advanced computing services for research centers at the State University of Moldova, partner institutions, and enterprises in the national economy. Currently, due to its uniqueness, the Laboratory has regional importance and can make a real contribution to enhancing collaboration between scientific centers in neighboring countries.

On the cluster, various classes of parallel algorithms are developed and implemented, lecture and laboratory courses are held in disciplines such as: parallel algorithms, parallel programming, etc. at the 1st and 2nd cycles of university studies, as well as within the framework of various research projects.

The HPC parallel cluster is equipped with licensed or “open source” parallel computing systems necessary for carrying out the necessary teaching and research processes. The available programming environments are OpenMPI, OpenMP, MPI, C, C++, F77, F90, ScaLAPACK, PETSc, Python, NumPy, MPI Python, MPI Java, Wolfram Mathematica.

At the moment, for the execution of parallel calculations, the HPC Laboratory offers 2 technical solutions of the HPC cluster type:

  • hpc.usm.md – 52 computing processors with a frequency of 2.5 GHz and a total memory of 68 GB RAM,
  • hpc2.usm.md – 100 computing processors, with a frequency of 2.6 GHz and a total memory of 272 GB RAM.

Contact details:
Hâncu Boris, dr. conf. univ.
boris.hancu@usm.md
hpc2.usm.md
hpc.usm.md