HP Visualization Cluster
An brief introduction to the remote visualization cluster at RZG
Hardware
- 1x login node viz00.rzg.mpg.de, 2x Intel E5540 quadcore CPUs, 144 GB RAM
- 5x standard visualization nodes (viz02-viz06), each equipped with 2x Intel W5580 quadcore CPUs, 144 GB RAM, 2x NVidia FX5800 GPUs
- 1x large visualization node (viz01), 4x Intel X7542 hexacore CPUs, 256 GB RAM, 2x NVidia FX5800 GPUs
- InfiniBand interconnect
- dedicated GPFS storage for
/uand/vizdata
File systems
There are two dedicated GPFS file systems for user data on the visualization cluster:
/u: File system for home directories (not backed up). Do not store large visualization datasets here!/vizdata: large scratch file system for visualization data (not backed up).
Furthermore, GPFS file systems from RZG's high performance computers are mounted, e.g.
/ptmp: scratch file system of the IBM Power6 machine.
AFS can be accessed from the visualization nodes (you may issue kinit first to obtain a token). For data security, copy your final visualization results to AFS or TSM archive regularly.
Software
- RZG's standard SLES 11 environment incl. modules
- various visualization and graphics applications (for details, see the quick start guide)
Accounts & application support
- To activate an existing RZG account on the visualization cluster, please send an informal email to visualization@rzg.mpg.de.
- Contact visualization@rzg.mpg.de and/or open a ticket in RZG's trouble ticket system.
How to use the visualization cluster
- The quick start guide provides step-by-step instructions on how to connect to the visualization cluster from Linux or Windows clients and gives information on how to work with visualization software.
- The reservation system VSRT is used to perform reservation and allocation of visualization resources.
