In the present page, we review the different compute parameters which can be set under the user interface.
The user can specify the maximum time limit for the calculation to be run for. The format is
HH:MM:SS in which H, M and S stand for day, hour, minute and second respectively. Thus, the example below would correspond to 12 hours exactly:
Please note that a calculation will be terminated when the runtime exceeds the maximum time specified in this field.
The user is offered the possibility to launch the desired job with a flexible set of hardware and resource allocation modes per queue. For example, the Debug ("D") queue is especially suited for preliminary tests, "fast" queues (eg. "OF")" are best for high-throughput or multi-node distributed memory runs, and "regular queues" (eg. OR) provide regular access to the scheduler.
Nodes / PPN¶
The desired number of computing nodes and the number of cores on each node (PPN = processors per node) can be selected, depending on the expected computational costs and requirements of the calculation under consideration. The necessary payments will have to be made as explained in this page, before the execution of any task can be made possible.
Finally, the user can be notified about the start of the calculation on the supercomputing cluster, about its termination, or about a possible accidental abortion.
The user can click on the button corresponding to each one of these Job statuses within the user interface to trigger the associated notifications. Alternatively, all three notification types can be activated simultaneously by clicking the user icon.
Further advanced options are offered in order to optimize the calculation. These depend on the simulation application.
The user should consult the corresponding application pages for more information about the advanced options which pertain to each respective application.