A Professional Approach to the Study of the Earth’s Interior

Computing capacity

Clusters

110-node cluster (HP DL360 G9 servers)

  • 2640 compute cores
  • total 21 TB RAM
  • NetApp RAID-DP disk array of 564 TB
  • Ethernet network infrastructure powered by HP FlexFabric 5700/5900 (10/40 Gbps)

8-node cluster (Dell PE M630 servers)

  • 224 compute cores
  • total 4 TB RAM
  • NetApp RAID-DP disk array of 62 TB
  • Ethernet network infrastructure powered by Dell Force10 MXL (10/40 Gbps)

Servers

5 Dell HPDL580 Gen9 Servers

  • 360 compute cores
  • 10 TB RAM
  • 150 TB disk array

HP DL580 Gen8 Server

  • 64 compute cores
  • 1 TB RAM
  • 48 TB disk array

Dell PowerEdge R930 Server

  • 64 compute cores
  • 2 TB RAM
  • 30 TB disk array

TOTAL

  • 3352 compute cores
  • 38 ТB RAM
  • 864 TB of disk storage
  • performance = 35 TFLOP according to Linpack

Development

  • Stage 1 (finished) - 3352 compute cores
  • Stage 2 (in progress) - up-scaling to 10,000 compute cores
  • Stage 3 (planned) - up-scaling to 14,700 compute cores