Cluster of the High Performance Computing Centre has the following structure.

System 1:

  1. Sites:
    1. 44 nodes with two Quad-core Intel Xeon E5440 @ 2.83 Ghz processors and 8 GB of RAM
    2. 68 nodes with two Dual-core Intel Xeon 5160 @ 3.00 Ghz processors and 4 GB of RAM
  2. Data exchange network: InfiniBand
  3. Disk space: 6 Tb distributed file system based on the LustreFS
  4. Capacity: 7, peak linpack TFlops TFlops 5.7
  5. ОС: CentOS release 5.2
  6. MPI: openmpi 1.2.8
  7. The local resource manager: Slurm
  8. Access protocol: SSH

System 2:

  1. Sites: 16 nodes with two Quad-core Intel Xeon E5345 @ 2.33 Ghz processors, 8 GB of RAM and disk capacity of 500 GB
  2. Data exchange network: InfiniBand
  3. ОС: MS Windows Server 2008 HPC Edition
  4. MPI: MS MPI 2.0.1551
  5. The local resource manager: HPC Job Manager
  6. Protocol access: RDP

Node motherboard supports IPMI interface for remote control.

Data storage system includes the following components:

  • 6Tb + 6.5 Tb of space on the base of FalconStor storage systems
  • 20 Tv space in strip archive

Cluster diagram

Схема кластера