Fluent/Batch

From ScientificComputing
Jump to: navigation, search

For submitting a fluent job, you need to create a journal file, which contains the commands that fluent should run. You can submit a fluent job for the journal file my_journal_file.jou with the following command.

bsub [LSF options] "fluent [fluent options] -i my_journal_file.jou"

Here you need to replace [LSF options] with LSF parameters for the resource requirements of the job. Please find a documentation about the parameters of bsub on the wiki page about the batch system. Note that for parallel jobs, it is sufficient to use the -n option of bsub. The batch system will automatically set the fluent parameter for specifying the number of cores, based on the value that you specify for -n. Please find below some description for the command line parameters, the are in the submission command specified as [fluent options].

Parameter Description
-cc Use the classic color scheme
-env Show environment variables
-mesher Start Ansys Fluent in meshing mode
-nm Don't display mesh after reading it
-post Start the Ansys Fluent post-processing-only executable
-rx Specify realease number

A typical Fluent job submission could look like the following:

bsub -W 8:00 -R "rusage[mem=2048, scratch=2048]" "fluent 2d -i fluent_case.jou > output.out"

In fluent, you can choose between single and double precision solvers. The choice can be specified via the parameter that indicates if the model is two or three-dimensional. In this example command, ouput is redirected to the file output.out.

Keyword Description
2d two-dimensional model single precision
2ddp two-dimensional model double precision
3d three-dimensional model single precision
3ddp three-dimensional model double precision