| |
Statistics
Horseshoe4
The Horseshoe4 cluster started production 21/8/2006.
After 3 years of hard work but stable production it was stopped 31/08/2009.
Conceptual overview
- 200x computing nodes (800 CPU cores, 55 TB user disk space)
- 9x 24-port gigabit ethernet switch in two stacks
- 1x front-end machine
- 1x infrastructure server
Hardware
Infrastructure
- Front-end Server
- Dell PowerEdge 2950, 2x 2,66Ghz Intel Woodcrest CPUs, 8 GB Ram, 2x160 GB HDD (7200rpm, 8 MB buffer, SATA150), 2x Gigabit Ethernet
- Infrastructure Server (batch, queue, web, name-services, mail)
- Dell PowerEdge 1650, 2x 1.26 GHz PIII, 512 MB Ram, 2x 36 GB HDD, 2x Gigabit
- Switches
- 9 D-link SR3324
- 2 D-link SRi3324
- The switches are organised in two stacks, each connected to the infrastracture switch with an 8 Gb/s LACP trunk.
Cluster nodes
- Computing nodes
- 160x Dell PowerEdge 1950 1U rackmountable servers with 2 2,66Ghz Intel Woodcrest CPUs, 4 GB Ram, 2x160 GB HDD (7200rpm, 8 MB buffer, SATA150), 2x Gigabit Ethernet
- 40x Dell PowerEdge 1950 1U rackmountable servers with 2 2,66Ghz Intel Woodcrest CPUs, 8 GB Ram, 2x160 GB HDD (7200rpm, 8 MB buffer, SATA150), 2x Gigabit Ethernet
Software
- Operating system on nodes and frontend
- OpenSuSE GNU/Linux 10.1 on a 2.6 kernel
- Batch- and scheduling software
- Torque 2.1.2
MAUI 3.2.6
- Compilers
- GCC 3.3.5
PGI Workstation 6.1
Intel C/C++ and Fortran 77/9x compilers v.9.1
- Applications and Libraries
- OpenMPI 1.1
- See Docs.
|
|
 |
