You can use the built-in CFX Batch job template on the Applications panel to run a CFX batch job. The template has been pre-configured with job submission settings that are specific to CFX, and works with all supported job schedulers. For general information about CFX solver runs, see Solver Run Overview in the CFX Solver Manager User's Guide. For detailed information about the solver run settings, see Define Run Dialog Box topic in the CFX Solver Manager User's Guide.
Click the Applications icon
on the title bar, and then click CFX
Batch on the Applications panel.
On the first screen of the job wizard, specify the tag to apply to the job (if desired), the job name, and the working directory. For details see Setting Up a Batch Job for details.
Click , then select the input files to be uploaded to the working directory. See Setting Up a Batch Job for details.
Click , then specify the CFX job submission settings:
Select the queue to which you want to submit the job.
Note: CFX jobs may fail if submitted to a queue that uses a local scratch directory. Ensure that you select a queue that uses a shared scratch directory.
The version of CFX that will be launched.
The available versions in this drop-down are
determined by the
solverVersions setting in
the ekm.xml file. For details
see Specifying Remote Process
Policies (<remoteProcess>) in
the Administration
Guide.
For CFX jobs, live job monitoring and control will not be available if the solver version is older than 17.1.
Click to select an input file from the working directory.
The settings on this tab correspond to settings on the Run Definition tab of the CFX Solver Manager’s Define Run dialog box. Refer to Run Definition Tab in the CFX Solver Manager User's Guide.
| Type of Run: Choose Full or Partitioner Only. When Full is selected, the Run Mode setting is available. To specify partitioner settings, enable the Show Advanced Controls check box, and then select the Partitioner tab. |
| Double precision. Check the box to select Double Precision. |
| Run Mode: During a Serial run, all computation is done by a single process running on one processor. A Parallel run divides computation into more than one process and is done on more than one processor in a single machine (local parallel processing) or on more than one machine (distributed parallel processing). |
| (Parallel) Number of partitions: The number of partitions should not exceed the number of available processor cores. For local parallel setup, you can select between 2 and 16384 partitions. |
When Parallel processing is selected, you must select a queue that does not use a local scratch directory. Distributed CFX jobs require a shared cluster directory that all compute nodes can access.
On Linux queues, when Parallel processing is selected, passwordless SSH access must be set up for each cluster node.
The settings on this tab correspond to settings on the Initial Values tab in the CFX Solver Manager’s Define Run dialog box. Refer to Initial Values Tab in the CFX Solver Manager User's Guide.

Enabling this check box displays additional Partitioner, Solver and Interpolator tabs.
On the Run Definition tab you can set the Message Passing Interface (MPI Type) to either Platform MPI or Intel MPI. Refer to Message Passing Interface (MPI) for Parallel in the CFX Solver Manager User's Guide for details.
The settings on this tab correspond to settings on the Partitioner tab in the CFX Solver Manager’s Define Run dialog box. See the Partitioning and Partitioner Tab sections in the CFX Solver Manager User's Guide for more information.

The settings on this tab correspond to settings on the Solver tab in the CFX Solver Manager’s Define Run dialog box. For details see Solver Tab in the CFX Solver Manager User's Guide.

The settings on this tab correspond to settings on the Interpolator tab in the CFX Solver Manager’s Define Run dialog box. For more information see Interpolator Tab in the CFX Solver Manager User's Guide.

If enabled, files in the job's working directory will be copied to a repository folder after the job has been executed.
Once enabled:
Click to select an Output folder for the files.
If you would like to import specific files or file types into the repository, specify a comma-separated list of patterns in the File patterns box, as shown in the example below. Otherwise, if you specify no patterns, all files in the working directory will be imported.

Click to execute the job. During the solution you can view live solution monitors, and stop the solution if necessary. See Working with CFX Server-Mode Jobs for details.