- Infos im HLRS Wiki sind nicht rechtsverbindlich und ohne Gewähr -
- Information contained in the HLRS Wiki is not legally binding and HLRS is not responsible for any damages that might result from its use -
Access and usage models: Difference between revisions
No edit summary |
|||
| Line 1: | Line 1: | ||
Please find more information in the online documentation of your selected system. | |||
== Access == | |||
Access to the HLRS compute servers is possible using following methods | |||
Grid access via Unicore within DEISA project | |||
or Globus Toolkit 4 (on the bwGRiD cluster only) | |||
Interactive access using ssh | |||
other protocols like telnet, rsh ... are not supported. | |||
Grid-only users may be restricted to Unicore/GT4 access only. To get full support we strongly recommend to file a project proposal. This includes interactive access via ssh, which is helpful for program development and porting tasks. | |||
How to get an account | |||
Revision as of 11:31, 30 March 2016
Please find more information in the online documentation of your selected system.
Access
Access to the HLRS compute servers is possible using following methods Grid access via Unicore within DEISA project or Globus Toolkit 4 (on the bwGRiD cluster only) Interactive access using ssh other protocols like telnet, rsh ... are not supported. Grid-only users may be restricted to Unicore/GT4 access only. To get full support we strongly recommend to file a project proposal. This includes interactive access via ssh, which is helpful for program development and porting tasks. How to get an account
Usage
In general three types of jobs are supported
- Interactive This jobs are intended for small tests and development tasks like debugging. Resources, like number of nodes, memory, cpu-time are very limited for this type of jobs.
- Batch Jobs This type of jobs is used for production runs
- Heterogenous Jobs This special type of batch jobs utilizes multiple type of execution nodes within a job. For example graphic nodes and cluster nodes.
Data Handling
The user HOME space is limited in terms of size and performance. Computational jobs run inside a Work_space. This Work_space is for active projects only and has a limited life time. If a project gets inactive for some reason, the user has to save its data to an archive system e.g. High Performance Storage System (HPSS).
Mailing List
to provide users with important information about HWW systems like announcements, problems, emergency maintenance, a mailinglist has been established:
Subscribe to the mailing list via web interface hwwsysnews or write an email to hwwsysnews-request@listserv.uni-stuttgart.de with the subject subscribe.