After you add a new job to your job repository you need to configure it for the execution in the grid. You need to describe the structure of the job, the applications you want to use and provide details on what the applications should actually do.
The configuration of each job starts with creating some tasks of your job. Each job must consist of at least one task. The first task created for a job is called the main task.
To create a new task click 'add' at the bottom of the 'Job main menu'. A new screen will appear.
The new screen contains a form built of six fields. The first one is the 'Task name' and this field must be filled. The 'Task description' field allows to describe the task in a few words. As with the job deascription, you are not required to provide a secription for a task of your job.
The third field is the 'Task type'. There are three task types: single, sequence and parallel.
The remaining three fields are active only when the you choose your task to be 'single'.
The first of these fields, the 'Task process count' refers to the number of processes that the task will be divided into during the execution. This number refers to mpi applications.
The second field, 'Task application', allows you to choose an application you want to use within your task. To get some more information about an application select it from the dropdown list and clicking 'Get application info'.
The last field, 'Task virtual application' allows to choose one of virtual applications. A virtual application is a defined structure of tasks allowing to quickly build a job with no toilsome task after task addition. A virtual application simply allows you to save much of your work!
Use these instructions to create as many tasks for your job as required. The only thing to remember is that a single type task cannot contain any subtasks. This means that a job for which the main task is single can contain exactly one task.
« Previous page | Tutorials index | Next page »