Page Content
| Number of
nodes | 35 |
| Processors (per
node) | 2x HexaCore-Xeon X5650 (2.67 GHz) |
| GPUs (only node291-node293) | je 2 nVidia
Tesla S2050 |
| Memory (per node) | 48GB
(GPU nodes) or 24GB |
| Network
(intern) | QDR-Infiniband + Gigabit
Ethernet |
| Bandwith | ~2500 MB/s
(MPI) |
| Network
(outwards) | Gigabit Ethernet |
| Local
disk (/scratch) | ~260
GB |
| Number of
nodes | 40 |
| Processors (per
node) | 2x QuadCore-Xeon X5550 (2.67 GHz) |
| Memory (per node) | 24GB oder
48GB |
| Network
(intern) | QDR-Infiniband + Gigabit
Ethernet |
| Bandwith | ~2500
MB/s (MPI) |
| Network
(outwards) | Gigabit Ethernet |
| Local disk (/scratch) | ~260
GB |
| Number of
nodes | 16 |
| Processors (per
node) | 2x QuadCore-Xeon X5550 (2.67 GHz) |
| GPUs | 14x 2 nVidia Tesla C1060 2x 4 nVidia GTX 295 |
| Memory (per
node) | 24GB |
| Network
(intern) | QDR-Infiniband + Gigabit
Ethernet |
| Bandwith | ~2500 MB/s
(MPI) |
| Network
(outwards) | Gigabit Ethernet |
| Local
disk (/scratch) | ~220
GB |
| Number of
nodes | 36 |
| Processors (per
node) | 2x QuadCore-Xeon X5550 (2.67 GHz) |
| Memory (per node) | 24GB oder
48GB |
| Network
(intern) | QDR-Infiniband + Gigabit
Ethernet |
| Bandwith | ~2500
MB/s (MPI) |
| Network
(outwards) | Gigabit Ethernet |
| Local disk (/scratch) | ~220
GB |
Technical data
The compute server can be used as a unified system, but the parts differ in terms of the age of the hardware, available hardware ressources and computing power.
The following tables show an overview about the Clusters.
When using the batch system, users have various options in choosing the appropriate hardware for their jobs.
Zusatzinformationen / Extras
Quick Access:
Schnellnavigation zur Seite über Nummerneingabe
Auxiliary Functions
This site uses Matomo for anonymized webanalysis. Visit Data Privacy for more information and opt-out options.
Copyright TU Berlin 2008